1 The New Child Labor: Digital Educational Privatization and the Contest over Students’ Bodies and Minds
Educational privatization has been aggressively expanding on two fronts. Traditional educational privatization, in the form of charter schooling and vouchers and scholarship tax credit neo-vouchers (Welner, Reference Welner2008), has continued to expand in the United States (Lieberman, Stanford, & Ifatusin, Reference Lieberman, Stanford and Ifatusin2024) and globally (Verger, Fontdevila, & Zancajo, Reference Verger, Fontdevila and Zancajo2016). Under the second Trump administration, a Republican Congress radically cut federal education funding for compensatory educational spending and civil rights programs. At the same time, it shifted funding and political power to states in order to expand voucher-based privatization (Schultz, Reference Schultz2025a). Vouchers publicly subsidize and incentivize the use of private, for-profit, and nonprofit schools, encouraging parents to opt out of public schools. They also have a long track record of worsening educational resource inequalities, intensifying racial segregation, and eroding public educational systems (Carnoy, Reference Carnoy2017). Traditional educational privatization thus shifts governance and control over schools and districts to private and for-profit entities. The second front, digital educational privatization, extends these dynamics beyond institutional governance and into the organized extraction of students’ pedagogical, affective, and cognitive labor. It reframes not only schools but the student as a site of value production.
Digital educational privatization has largely operated within schools, acting on students, teachers, curriculum, pedagogy, and school cultures to introduce profit-seeking technologies into everyday educational practice. Digital privatizers seek lucrative contracts with schools and districts or strategically offer free platforms in exchange for the right to capture commercially valuable student and teacher data (Sadowski, Reference Sadowski2019; Williamson, Reference Williamson2017; Perrotta et al., Reference Perrotta, Gulson, Williamson and Witzenberger2020; Selwyn, Reference Selwyn2016). They also widely use that data to market fee-for-service products and subscription upgrades to parents (Saltman, Reference Saltman2018). Ninety-six percent of school apps send data to third parties including advertisers (Internet Safety Labs, 2022) and 75 percent of popular educational apps profit from student data (Common Sense Media, 2023). Digital educational privatization shifts pedagogy onto screens and frames technology and technology companies, rather than educators, as the real teachers. It falsely purports to be culturally relevant and personalized even as it delivers highly standardized, scripted, and homogenized content. Digital educational privatization makes incessant data production and testing the dominant mode of pedagogy and systematically undermines teachers’ professional capacities to relate knowledge and learning to student experience, community knowledge, and the broader social and political world (Selwyn, Reference Selwyn2019).
As I have discussed in The Disaster of Resilience (Saltman, Reference Saltman2023), digital educational privatization has expanded in part by using Social and Emotional Learning (SEL), mindfulness, and meditation programs as curricular content that introduces technology into schools and extracts valuable data from students. Under a therapeutic and humanitarian guise of attentiveness to feelings, these programs largely aim to instill disciplinary self-regulation, conformism, and the internalization of blame for social conditions not of students’ making (Slater, Reference Slater, Saltman and Nguyen2022; Manolev, Sullivan, & Slee, Reference Manolev, Sullivan and Slee2019). Despite the language of resilience, they fail to provide students with the tools to comprehend their emotions in relation to the social forces, interests, and ideologies that produce them (Means, Reference Means2016; Saltman, Reference Saltman2023).
If traditional educational privatization through charters and vouchers operates predominantly at the macro level, targeting schools, districts, and state systems, digital educational privatization operates predominantly at the micro level, targeting the bodies, affects, and consciousness of students and teachers. Traditional educational privatization has long operated at the micro level as well. As part of neoliberal ideology, it sought to reimagine the inner workings of schools through corporate culture, casting students and parents as educational consumers, positioning knowledge as a deliverable commodity, framing school culture on the model of business, and representing students as incipient entrepreneurs. What must not be missed, however, is that digital educational privatization makes the bodies, affects, and perceptual attention of students and teachers the very locus of accumulation. It gets eyeballs on screens, converts attention into clicks, and transforms pedagogical activity itself into data production.
Much, if not most, criticism of digital educational privatization has focused on student data privacy and concerns about surveillance. Activists, organizers, and scholars have rightly expressed alarm over the uses that businesses make of the data students generate through educational technologies. Some activists have successfully enacted legislation to limit these practices. For example, Illinois Families for Public Schools successfully lobbied the Illinois State Legislature to expand student digital privacy protections. Critics of digital surveillance and surveillance capitalism link these concerns to broader critiques of school commercialism, managerial control, and repression. In “The datafication of discipline,” Manolev, Sullivan, and Slee (Reference Manolev, Sullivan and Slee2019) show how the widely adopted app ClassDojo technologizes repressive and behaviorist disciplinary practices. Shoshana Zuboff, in The Age of Surveillance Capitalism (Reference Zuboff2019), warns that technology corporations do not merely infringe on privacy but actively acquire personal data to shape desires, conduct, and consciousness, deploying extracted data as a form of asymmetrical power over users. The Cambridge Analytica scandal, in which harvested data was used to target political advertising during the 2016 US elections, brought these practices into wider public view. As important as critiques of surveillance capitalism are, they remain distinct from and insufficient for grasping how capitalist relations of production are introduced into schooling through digital educational privatization itself. At stake is not only surveillance but the reorganization of students’ and teachers’ everyday pedagogical activity as a form of value-producing labor (Robinson, Reference Robinson2024). It must be stressed that digital child labor is not experienced uniformly. The datafication and extraction processes described here are structured by existing racial and gender hierarchies, shaping who is most intensely monitored, evaluated, rewarded, or punished through educational processes producing very different academic and life chances. While this work does not develop a full account of these dynamics, race and gender must be understood as integral to how digital educational privatization and child labor are organized and lived.
In what follows, I argue that digital educational privatization introduces mandatory technologies into schools that compel students to participate in the production of commercially valuable data for business. In this sense, digital educational privatization constitutes a new form of child labor that has yet to be widely recognized or named as such. The first part of this section examines the political economy of data extraction and value production in the new child labor. The second analyzes the cultural politics involved in its formation and normalization. The third situates the new child labor within broader political, economic, and cultural trends and antagonisms, including the expansion of authoritarian governance, the private-sector pillage of the public sector, and the systematic undermining of public institutions as foundations for democratic culture, agency, and struggle (Brown, Reference Brown2019; Harvey, Reference Harvey2005; Giroux, Reference Giroux2004). I conclude by outlining what is required to recognize and challenge the resurgence of child labor, both new and old, and to foster critical and democratic digital culture as part of a broader effort to reclaim public education, recover democratic values and practices, and confront rising authoritarianism on a global scale.
1.1 The Political Economy of the New Child Labor
Critical scholars recognize that data functions as a commodity and that users are engaged in the manufacture of this commodity (Sadowski, Reference Sadowski2019). It is widely understood that the data produced through the use of touchscreen and networked digital devices is captured, sold, bought, and circulated through hundreds of corporations and state actors (Birch, Cochrane, & Ward, Reference Birch, Cochrane and Ward2021; Sadowski, Reference Sadowski2019). As Jathan Sadowski argues, data is lucrative not only because of its immediate uses but because of its speculative potential for future applications, even when those uses are not yet known. For this reason, businesses aim to accumulate data as capital. Whether processed through large-scale analytics to generate profitable knowledge or simply held, bundled, traded, or sold, data retains exchange value independent of any immediate application.
Corporations therefore have strong financial incentives to get programs and platforms into schools, where the routine activities of students generate continuous streams of data that can be captured and accumulated. As Williamson (Reference Williamson2017) demonstrates, contemporary educational platforms embed data-producing activity directly into public schooling itself. Technology, media, and education corporations seek profit by entrenching such activity within everyday educational practice, effectively putting students to work generating data. Some of these companies sell instructional services and technologies directly. Many others offer free platforms that secure access to schools while retaining the right to use student-generated data to market fee-for-service products and subscription upgrades to parents (Williamson, Reference Williamson2017). This model operates in addition to, not instead of, the extraction of data as capital. The primacy of data extraction as a revenue strategy renders the educational value of many digital programs and products structurally secondary to their economic function (Selwyn, Reference Selwyn2019), a problem I return to in the section on cultural politics.
A number of scholars (Sadowski, Reference Sadowski2019; Fuchs, Reference Fuchs2014b; Jhally, Reference Jhally1987) have noted the similarity between contemporary data-extraction models and what mass communication theorist Dallas Smythe (Reference Smythe1981) famously described as the “audience commodity.” Writing in the era of network television, Smythe argued that the audience itself was the product produced through television viewership and sold to advertisers. Broadcast content functioned primarily to capture attention and deliver audiences to advertising. The uncompensated labor of learning consumer products and learning consumerism through advertisements was the price viewers paid for access to content. In this formulation, learning itself functioned as unpaid labor. Smythe emphasized that the commodity being sold was not the consumer goods featured in advertisements but the viewers, who were sold by networks to advertisers.
Internet commercialization operates in a structurally similar manner to network television, in that advertising continues to drive much online content (Fuchs, Reference Fuchs2014a). Today, internet users perform the uncompensated labor of learning advertisements and the uncompensated labor of producing data through clicking, scrolling, and interacting with platforms. Increasingly, students in schools are compelled to perform analogous uncompensated labor by generating commercially valuable data through the mandatory use of digital educational programs and platforms (Molnar, Reference Molnar2019).
Contemporary arrangements of digital child labor extraction in schools can be understood, in some respects, as both an extension and a transformation of the long historical legacy of school commercialism. School commercialism seeks to insert advertising into schools and market to youth as a captive audience. Like the audience commodity, it depends on the audience performing the labor of learning consumption. School commercialism expanded with the rise of neoliberal ideology in the early 1980s (Saltman, Reference Saltman2000). During the 1980s and 1990s, it took the form of junk food advertisements in textbooks and the introduction of television advertising into classrooms, most notably through Christopher Whittle’s Channel One program, which provided brief and superficial news content in order to expose students to advertisements for clothing, cosmetics, and other consumer goods (Saltman, Reference Saltman2005).
The “cola wars” of the 1990s saw Coca-Cola and Pepsi competing for exclusive school contracts that placed vending machines in schools and corporate billboards on school buildings and athletic fields (Saltman, Reference Saltman2000). Oil companies such as BP Amoco and Shell similarly marketed their brands through sponsored science curricula, among other examples (Goodman & Saltman, Reference Goodman and Saltman2002). Technology companies provided hardware to schools to promote their products, while heavy industries such as Toyota and BMW, with the assistance of the National Association of Manufacturers, developed corporate-sponsored curricula under the rubric of Career and Technical Education (CTE) (Bryant, Reference Bryant2020). While some of these initiatives aimed to introduce publicly sponsored private worker training, most school commercialism in the postwar era sought to market brands to students and cultivate long-lasting positive brand impressions among youth (Linn, Reference Linn2004; Schor, Reference Schor2004; Molnar, Reference Molnar2005).
Through the theoretical lens of social and cultural reproduction (Bowles & Gintis, Reference Bowles and Gintis2011), school commercialism can be understood as a form of labor insofar as schools taught knowledge and skills embedded within work ideologies, preparing students for their future roles in the capitalist economy. This long-term investment in producing knowledgeable, and more importantly docile, future workers was time- and labor intensive. As Nancy Fraser (Reference Fraser2002), Zygmunt Bauman (Reference Bauman1998), and others have argued, the transition from the Fordist industrial economy to the post-Fordist neoliberal service economy coincided with transformations in self-regulation and social control. Post-Fordist modes of control increasingly relied on more direct, immediate, and corporeal forms of bodily regulation (Fraser, Reference Fraser2002).
The time-intensive work of cultivating learned self-discipline in schools was increasingly supplemented, and in many cases displaced, by heightened repression. This shift took the form of expanded security apparatuses, zero-tolerance policies, and the prisonization and militarization of schools (Saltman & Gabbard, Reference Saltman and Gabbard2010), alongside the rapid growth of pharmaceutical instruments for managing behavior and affect, including amphetamines for attention and diazepines for anxiety (Berardi, Reference Berardi2015; Saltman, Reference Saltman2025b). In this context, school commercialism became less about long-term investment in future profit through the slow cultivation of capitalist subjectivity and more about immediate profiteering through contracts with districts and schools. These contracts ranged from school-management services to SEL and behavior-management programs to digital learning products and systems for tracking student behavior.
School commercialism thus constituted a significant component of ideological education in a transforming economy and society that increasingly positioned schooling for work and consumerism while decreasingly positioning it as a site of humanistic development or democratic and political preparation. Digital child labor, however, differs in important ways from both the legacy of school commercialism and the historical role of schooling in reproducing the labor force.
Digital child labor is not organized primarily around advertising, consumption, or ideologies of consumerism, nor around long-term investment in future labor. It is organized around immediate production and value extraction. Through the mandatory use of digital programs, students produce data commodities in real time as they engage in routine educational activity. Digital child labor must therefore be understood as part of a global commodity chain that begins with child labor in peripheral nations and extends to child labor in core nations, where students’ compulsory school activity is reorganized as value-producing work (Fuchs, Reference Fuchs2014a; Crary, Reference Crary2022).
As Jonathan Crary argues in Scorched Earth (Reference Crary2022), the use of digital products in wealthy nations depends upon the extraction of minerals in poorer ones, often through super-exploited forms of child labor. Cobalt and coltan used in lithium-ion batteries that power digital devices, for example, are mined extensively by children in the Democratic Republic of Congo (Amnesty International, 2016). Crary emphasizes that the techno-utopian ideology framing increasing digitalization as ecologically sustainable is demonstrably false. As Kate Soper argues in Post-growth Living (Reference Soper2020), the environmental devastation produced by extractive mining, combined with the intensive reliance on fossil fuels required to support the rapid expansion of server farms and AI infrastructure, demonstrates that increased digitalization does not tend to produce either greener ecological outcomes or reductions in labor exploitation. The expansion of digital educational privatization thus represents a largely unrecognized dimension of intensified labor exploitation, operating simultaneously through digital software in schools and through the material labor and environmental devastation embedded along the commodity chains of digital hardware.
As Soper and ecological economists emphasize, the expansion of so-called green technologies entails significant increases in extractive mining and continued fossil fuel combustion to produce electric vehicles, solar panels, batteries, and wind turbines (Hickel, Reference Hickel2020). The capitalist promise of unlimited growth collides with the reality of a finite planet. The only viable direction toward a genuinely sustainable future, and perhaps any future at all for youth, requires a significant reduction in production and consumption. Such a shift demands not merely technological substitution but a revaluation of collective cultural priorities and a transformation of education capable of cultivating a culture of sustainable living rather than green-capitalist substitution.
In Cannibal Capitalism (2022), Nancy Fraser argues that contemporary capitalism increasingly consumes its own conditions of social reproduction, undermining the very institutions and practices upon which it depends. This dynamic characterizes both traditional educational privatization and digital educational privatization. As public goods such as public schools are dismantled to create private markets, this process hollows out the reproductive, democratic, and social functions of schooling itself. Despite overwhelming evidence that voucher schemes erode educational quality (Carnoy, Reference Carnoy2017), they continue to be promoted and expanded in the interests of capital accumulation and in the service of shrinking and delegitimizing the public sphere. Contemporary capital cannibalizes its own conditions of reproduction in part through raclalized expropriation of labor and natural resources on a global scale. This is glaringly evident in the Global North as digital child labor most intensively targets working-class and poor black and brown students for data production. Meanwhile, in the Global South black child labor in mines is expropriated for mineral extraction for tech hardware.
In the United States, this tendency for racialized expropriation has been exemplified through the dismantling of the Department of Education, the defunding of public schools, and the redirection of educational spending toward voucher subsidies under Trump (Schultz, Reference Schultz2025a). Digital educational privatization, despite extensive evidence documenting the poor quality of many programs and platforms (Selwyn, Reference Selwyn2019; Slama et al., Reference Slama, Pane, Steiner, Hamilton and Pane2019; WWC, 2022; Williamson, Reference Williamson2017), continues to expand due to the combined force of profit imperatives and the legitimating ideologies of techno-utopianism and neoliberalism (Means, Reference Means2018).
Techno-utopianism frames social problems and progress as best, and often only, addressed through technological solutions. Neoliberal ideology reduces social life to market values and metaphors and installs audit culture as a dominant mode of governance. Neoliberalism has a strong affinity with techno-utopianism, as both require that all values be rendered quantifiable, measurable, and comparable in the service of commercial extraction and market expansion. Technoauthoritarians such as Peter Thiel, Elon Musk, Marc Andreessen, Balaji Srinivasan, Alex Karp, and others deepen this fusion by seeking to naturalize inequality and hierarchy, often invoking claims about biologically differentiated capacities across races and genders (Slobodian, Reference Slobodian2023, Reference Slobodian2025). Drawing on discredited conceptions of intelligence and retrograde forms of eugenic reasoning, these tech oligarchs frame technological acceleration as inherently beneficial regardless of its human, political, or environmental costs (Slobodian, Reference Slobodian2023, Reference Slobodian2025).
As tech oligarchs develop and sell AI-driven weapons systems and expansive digital surveillance infrastructures that monitor and discipline populations, they simultaneously advocate deep cuts to public spending on the caregiving functions of the state and radical deregulation of capital, prioritizing profit over public interest (Zuboff, Reference Zuboff2019). These tech oligarchs explicitly acknowledge the incompatibility of democracy and capitalism and reject democracy as an impediment to technological and market acceleration (Slobodian, Reference Slobodian2025; Klein & Taylor, Reference Klein and Taylor2025).
They often frame their embrace of repression, the destruction of democratic institutions, and environmental devastation as sacrifices undertaken on behalf of hypothetical future populations imagined to benefit from unchecked technological expansion. This framing obscures the present reality in which such expansion overwhelmingly enriches oligarchs while imposing social, political, and ecological costs on everyone else. It would be a mistake, however, to treat the actions and statements of tech oligarchs as merely the maneuvers of an elite cabal of bad actors. They must instead be understood systemically, as expressions of a changing global class structure in which tech oligarchs occupy the upper strata of the transnational capitalist class (Robinson, Reference Robinson2024).
In Into the Tempest, William I. Robinson (Reference Robinson2019) argues that the expansion of repression in education must be understood in relation to the material and ideological projects of the transnational capitalist class. For Robinson, the global agenda of educational privatization and increasingly repressive schooling reflects capital’s need for more intensive, immediate, and corporeal forms of social control. Repressive pedagogies, scripted instruction, and the standardization and homogenization of curriculum, alongside expanded security apparatuses, express growing demands for obedience and compliance in workplaces organized under financialized capitalism. As automation expands and employment opportunities contract, schools serving poor and working-class communities increasingly function as warehousing institutions, feeding into the carceral system and the repressive apparatus of the military.
Robinson argues that as capital confronts crises of reproduction, continued accumulation increasingly depends on what he terms militarized accumulation, in which capital actively disrupts existing social formations to open new sites for profit-making (Robinson, Reference Robinson2019). This process includes the targeting of previously noncommercialized domains such as subjectivity, affect, and the lifeworld. Digital educational privatization exemplifies this dynamic by reorganizing students’ compulsory activity, energy, and subjectivity into sites of value production through data extraction. In this sense, schooling itself is transformed into a site of militarized accumulation, where value is produced through compulsion, dispossession, and the forced incorporation of children into digital markets.
The student’s body is a central site of material struggle among competing class agendas. As the transnational capitalist class extracts uncompensated labor from children in the form of data production, the question arises of what working-class interests look like under these conditions. One possible response would be to popularize recognition that children are producing commercially valuable data and to argue that the economic value generated should be realized primarily by the worker, that is, by the child. Under such an arrangement, children would be financially compensated for the value of the data they produce.
Alternatively, the market value of student-generated data could be captured and redirected not to individual children but back to schools or districts. Either approach would undermine the predatory business models of for-profit digital app producers by stripping them of their ability to profit from embedding their products in schools. Such arrangements would also limit companies’ capacity to use student data to market fee-for-service products to parents. This strategy has the advantage of directly challenging labor exploitation and undercutting the profit motive driving digital privatizers to enter schools with dubious products. It would also recoup value created by children’s activity. At the same time, it carries an obvious drawback. It risks normalizing child labor in its new digital forms and reinforcing education primarily as an instrument of capitalist production rather than as a foundation for creative democracy (Dewey, Reference Dewey and Boydston1988), humanistic values, and the cultivation of a compassionate, caring, rational, critically reflective, and humane society.
Another seemingly straightforward solution would require that any data produced by students remains private or be immediately purged so that it cannot be transformed into a commodity by for-profit app producers. This solution, however, presents a serious problem. Data governance would remain in the hands of the same corporate actors whose business models depend upon extracting, circulating, and commercializing that data. Tech philanthrocapitalist organizations such as the Chan Zuckerberg Initiative, Emerson Collective, and the Omidyar Network have repeatedly demonstrated that they cannot be trusted to police themselves (Saltman, Reference Saltman2018). Their founders have structured these organizations as limited liability companies (LLC), insulating them from public oversight and enabling opaque movement of data and capital between for-profit and nonprofit entities housed within the same corporate structure.
This organizational form has made it impossible to track, for example, the flow of data and money between the nonprofit Summit Learning platform and for-profit firms such as Byju’s, both housed within the Chan Zuckerberg Initiative. It also obscures the circulation of personnel, resources, and data between these initiatives and external corporations such as Facebook, whose engineers contributed to building Summit (Saltman, Reference Saltman2018). Philanthrocapitalist actors entering digital educational privatization have clearly learned from earlier experiments in charter-based privatization. Zuckerberg’s chartering of the Newark, New Jersey public schools could not fully conceal scrutiny of financial flows because charter schools are required to file publicly accessible tax documents (Russakoff, Reference Russakoff2015). Charter schools have long claimed public status when demanding funding and subsidies while asserting private status to shield operations, expenditures, and decision-making from democratic oversight. The LLC form circumvents accountability requirements that apply to public or nonprofit educational institutions, allowing philanthrocapitalists to move data, capital, and personnel opaquely across organizational units.
As voucher-based privatization expands in the United States, for-profit schools and education companies will predictably seek to increase profits by not only pursuing profit from the data commodity produced by the new child labor but also by reducing their largest expense: labor costs (Molnar, Reference Molnar2019). This dynamic will accelerate the replacement of teacher labor with technology. Voucher-based privatization will also incentivize homeschooling arrangements that allow families to retain public education funds while relying heavily on digital platforms for instruction. In doing so, it further enables the public subsidization of private, and often religious, education (Carnoy, Reference Carnoy2017).
1.2 Cultural Politics of the New Child Labor
The return of child labor must be understood in relation to historical efforts to eradicate child labor and institutionalize schooling as its alternative. In the nineteenth century, the common school movement sought to remove children from agricultural and industrial labor and relocate them into schools (Tyack, Reference Tyack1974). For Horace Mann, this project combined the cultural assimilation of immigrants, the political formation of citizens, the moral formation of human beings, and long-term investment in the reproduction of the labor force. It was also explicitly aimed at diffusing class antagonisms and creating the possibility of meritocratic uplift (De Lissovoy, Means, & Saltman, Reference De Lissovoy, Means and Saltman2016).
In this respect, the origins of public schooling share important affinities with Andrew Carnegie’s stated rationale for philanthropic support of libraries, universities, museums, and other cultural institutions. In The Gospel of Wealth (Reference Carnegie1899), Carnegie argued that ruling-class charity could forestall radical political movements, particularly socialism, that threatened to redistribute control over capital from the few to the many. He advocated expanding educational and cultural institutions as a means of individualizing responsibility for social advancement. This strategy shifted the burden of mobility away from structural inequality and onto personal effort. If everyone could be educated and “cultured” through access to libraries, museums, and schools, then social advancement could be framed as the outcome of individual merit rather than collective struggle.
It is worth noting here that late twentieth- and early twenty-first-century venture philanthropy played a crucial role in preparing the ideological terrain for digital educational privatization. As I argue in The Gift of Education (Reference Saltman2010), foundations such as Gates, Walton, Broad, and Dell leveraged philanthropic capital to reshape public education governance through a neoliberal imaginary that framed schools as competitive private enterprises, parents as private consumers, and educational outcomes as matters of individual responsibility rather than collective provision. Gates backed charter school and Common Core expansion; Walton backed voucher expansion; Broad bankrolled database tracking and the corporatization of leadership. This project extended the older philanthropic strategy of depoliticizing inequality by shifting attention away from structural conditions and toward individualized performance, choice, and competition.
This ideological project of meritocratic individualization, championed by the robber barons of the industrial era and radicalized by the venture philanthropists, not only persists but is intensified through digital educational privatization. Students are framed as individually accountable for success or failure, while automated assessment and placement systems, including individualized data profiles, adaptive pathways, and algorithmically mediated assessments, are naturalized as impartial technologies rather than political instruments of sorting.
Today, many of the programs and platforms that introduce digital products into classrooms rely heavily on resilience-oriented concepts such as mindfulness, growth mindset, Social and Emotional Learning, grit, and meditation. As I have detailed in The Disaster of Resilience (Saltman, Reference Saltman2023), these concepts frequently constitute the core content of for-profit educational technologies. Like their nineteenth-century predecessors, contemporary resilience frameworks place responsibility for educational uplift squarely on the student, who is framed as deficient in self-control and therefore in need of behavioral and affective remediation delivered through digital products.
Within this discourse, human capital theory casts learned self-regulation, cultivated through behavior- and affect-targeting interventions, as the primary means of overcoming the “trauma” associated with poverty (Slater, Reference Slater, Saltman and Nguyen2022). Poverty-induced trauma is said to obstruct learning by interfering with students’ receptivity to instruction, while resilience programs promise to remove this obstruction and render students capable of absorbing sanctioned knowledge. In this way, contemporary resilience initiatives reproduce long-standing meritocratic assumptions embedded in public schooling, misframing structural inequality as an individualized problem of affect, behavior, and disposition. Students’ routine engagement with these digital products not only instills these ideologies but also continuously generates commercially valuable data, binding ideological education and uncompensated digital labor into a single pedagogical process.
Digital educational privatization is not only a terrain of material interests involving contracting, accumulation, and labor extraction. It is also a central terrain of cultural politics. Digital apps largely reproduce and intensify long-standing repressive approaches to teaching and curriculum, reshaping how knowledge is defined, delivered, and authorized in schools. AI-based reading programs such as Amira Learning, for example, replace teachers’ pedagogical judgment with scripted and automated instruction (Amira Learning, Reference Learning2023). So-called personalized learning programs similarly claim responsiveness to student experience, yet in practice operate through highly standardized, homogenized, and transmissional forms of instruction that deliver the same curriculum through individualized interfaces (Williamson, Reference Williamson2017; Cuban, Reference Cuban2018; Eubanks, Reference Eubanks2018). Despite the label, personalized learning is structurally incapable of addressing the relationship between knowledge and lived student experience, between knowledge and social context, or the ways knowledge becomes meaningful in relation to broader social forces, structures, and systems.
Social media–style classroom management apps such as ClassDojo and GoNoodle further illustrate how cultural politics operates through digital pedagogy. These platforms function primarily as tools of digital surveillance and behavior management (Manolev, Sullivan, & Slee, Reference Manolev, Sullivan and Slee2019) while simultaneously claiming to cultivate social and emotional regulation. Students are required to register, monitor, and manage emotions but are not provided conceptual resources for understanding the social, political, and material conditions that produce those emotions. Instead, such programs emphasize breathing, visualization, and self-soothing techniques that encourage students to manage, suppress, or ignore feelings rather than critically interpret them (Saltman, Reference Saltman2023; Means, Reference Means2016). Across these platforms, pedagogy is reduced to regulation, compliance, and affect management rather than inquiry, interpretation, or meaning-making.
Digital educational privatization programs therefore foreclose the conditions for inquiry, investigation, dialogue, debate, dissent, and curiosity, capacities that enable students to question, interpret, and produce meaningful knowledge (Selwyn, Reference Selwyn2016). The labor of producing data and decontextualized information comes at the expense of the co-production of knowledge through dialogic exchange, interpretation, and judgment. As Stuart Hall (Reference Hall1997) and Paulo Freire (Reference Freire1972) argued, knowledge and culture are not delivered intact but are produced relationally through dialogue, struggle, and interpretation.
By contrast, digital educational products overwhelmingly embrace a transmission model of pedagogy (Giroux, Reference Giroux1983), in which knowledge is presumed to be deposited into students (Freire, Reference Freire1972). This view reifies knowledge as static, settled, and external to interpretation while concealing the interests, ideologies, and social positions that shape claims to truth. Such assumptions stand in direct tension with an understanding of cultural politics, which recognizes that culture and knowledge are saturated with power relations and structured by contestation. Cultural politics holds that meanings are always subject to interpretation and debate yet tend toward what Hall (Reference Hall1997) described as “preferred meanings” as a result of shared assumptions, institutional arrangements, and dominant power relations.
As a conceptual framework, cultural politics presumes that meanings are contested and that those who advance particular meanings do so from material and symbolic positions of interest that shape meaning-making itself (Saltman, Reference Saltman2025a). Applied to digital educational privatization and the forms of child labor it produces, this perspective reveals how technologies students are compelled to use actively assert particular meanings while simultaneously closing down signification. These systems misrepresent the knowledge they confer as neutral, objective, and beyond interpretation or debate. They rest on a positivist epistemology that treats truth as a collection of decontextualized facts detached from their conditions of production, including the ideologies, social relations, and power structures that inform claims to truth (Adorno, Reference Adorno1999; Giroux, Reference Giroux1983; Saltman, Reference Saltman2022).
The denial of the cultural politics of knowledge embedded in many digital pedagogies carries with it a specific conception of the relationship between knowledge and politics. Democratic societies require citizens equipped not only with technical or instrumental knowledge for collective self-governance but also with democratic dispositions, including capacities for dialogue, debate, dissent, and critical judgment (Mouffe, Reference Mouffe2013). Public schools ideally model and enact democracy through dialogic forms of teaching and learning that cultivate curiosity and creativity and that provide students with intellectual tools for situating claims to truth in relation to authority, interests, and power relations. Pedagogy is never politically neutral. It functions as a formative practice shaping how students understand knowledge, authority, and democratic participation.
Digital programs, however, largely frame knowledge as static, inert, or dead rather than as a dynamic process produced through dialogic exchange (Selwyn, Reference Selwyn2016). Children are positioned as passive recipients who purportedly absorb knowledge through interaction with programs. This framing extends the long legacy of positivist and transmissional approaches that Paulo Freire famously criticized as “banking education” (Reference Freire1972), in that technology is presumed capable of delivering knowledge ever more efficiently by displacing the allegedly inefficient human teacher with a technological substitute.
Such a conception of knowledge stands in direct tension with theories of knowledge formation that understand knowledge as socially produced through dialogue among actors situated within unequal relations of power. As Freire, Hall, Bakhtin, Giroux, and other theorists particularly from the traditions of poststructuralism, pragmatism, and critical theory have argued, knowledge is dynamic, relational, and socially constructed, and must be understood in relation to the values, assumptions, ideologies, and material interests that undergird and inform claims to truth. The transmissional view instead presents knowledge as neutral, universally valuable, and disinterested. Techno-utopian frameworks intensify this tendency by treating knowledge not only as static but as dead, concealing the interests behind claims to truth while reducing knowing to a transactional activity oriented toward academic rewards rather than self-understanding, social interpretation, or collective agency. Whereas transmissional models transform education into an instrument of discipline, social control, and submission to the existing social order, dialogic and constructivist approaches position education as a foundation for democratic capacity, collective self-governance, and critical social transformation.
1.3 Situating the New Child Labor
The recent expansion of child labor in the form of digital educational privatization must be understood in relation to contemporary economic, political, and cultural forces. In the wake of far-right anti-immigrant policies and mass deportations in the United States, several states have turned to child labor as a source of cheap, highly exploitable labor. This shift is widely understood as a response to labor shortages produced by the removal of migrant workers under Trump-era deportation regimes (Gonzales, Sigona, & Zetter, Reference Gonzales, Sigona and Zetter2018). Florida, for example, has advanced legislation explicitly aimed at expanding child labor. “One proposal would allow sixteen-year-olds to work beyond the current thirty-hour-per-week cap, including longer days and later hours, even on school nights. Another would loosen restrictions on fourteen-year-olds working if they are homeschooled, enrolled in virtual education, or already graduated” (Wood, Reference Wood2025). This legislation, particularly its explicit targeting of virtual and homeschooling arrangements, exposes the relationship between the political right’s project to expand educational privatization and its renewed embrace of child labor. The right’s support for child labor must be understood as part of a broader class project that seeks to dismantle the public and caregiving functions of the state while privatizing public goods such as schooling (Bourdieu, Reference Bourdieu1998). As the Florida proposals illustrate, voucher-incentivized homeschooling creates conditions for the super-exploitation of children’s labor while simultaneously depriving them of meaningful educational opportunities.
The long-standing agenda of the political right has been to transform public education into a private industry. The aggressive dismantling of federal agencies, mass firings, and program closures under the Trump administration in 2025, including the gutting of the Department of Education, represents an effort to redirect public funding away from redistributive and compensatory programs aimed at mitigating class and racial inequality and toward state-level privatization schemes (Office of Management and Budget, 2025). This agenda extends a longer project to roll back social protections developed through the New Deal, the Civil Rights Movement, and the Great Society.
The contemporary political right is actively enabling the expansion of child labor by deregulating labor protections, weakening regulatory agencies, and promoting state-level rollbacks that allow minors to work longer hours and enter more dangerous industries (Economic Policy Institute, 2023). At the same time, right-wing economic and education policies, including the defunding of public schools, the suppression of adult wages, attacks on social safety nets, and the expansion of youth apprenticeship pipelines, intensify precarity and normalize early workforce participation as a moral or economic necessity. These developments are reinforced by ideological narratives that frame youth labor as character-building and by digital workforce and education products that integrate children directly into corporate labor streams. Together, these material and cultural conditions create a political-economic environment conducive to the expansion of child labor.
The renewed turn toward child labor is thus consistent with the political right’s broader effort to reverse a historically significant achievement: the legal and cultural displacement of children’s labor by schooling. That achievement, forged through late nineteenth- and early twentieth-century struggles, established education as the normative alternative to child labor.
The political right’s embrace of child labor also aligns with the global ascent of authoritarian and fascist movements seeking to forge a new hegemonic ideology capable of supplanting neoliberalism. This emergent formation advances a platform often summarized as the “three F’s”: Faith, Family, and Freedom. It seeks to fuse White Christian nationalism with homophobic and transphobic projects that reassert the heterosexual family as the core social unit, alongside an aggressively marketized conception of economic freedom (Butler, Reference Butler2024; Brown, Reference Brown2019). Put bluntly, the “three F’s” merge market fundamentalism with Christian fundamentalism. The political right embraces voucher-based privatization to pursue these ideological pillars: by shifting educational responsibility from the public sphere to the family, by enabling publicly funded religious and homeschooling arrangements, and by opening public education budgets to private investors who extract wealth through for-profit schools and education companies.Footnote 1
The resurgence of child labor enables an additional layer of accumulation in the form of low-paid, weakly regulated, or entirely unregulated youth labor (Economic Policy Institute, 2023). This retrograde effort to roll back social protections proceeds alongside a distinctly futuristic project that places children in front of screens and enrolls them in AI-driven education programs. These programs reorganize schooling as continuous data production, transforming children themselves into engines of value extraction. In this convergence of reactionary social policy and technological acceleration, digital child labor becomes a central mechanism through which authoritarian cultural politics and capitalist accumulation are mutually reinforced.
2 The Teacherless School: Alpha and the Rise of the AI School
Alpha School, a for-profit company, has launched what it claims is the first chain of schools in which AI replaces teachers for all core academic instruction (Alpha School, 2024a).Footnote 2 Alpha began as a private school in Austin, Texas, charging approximately $40,000 in annual tuition. The model organizes the school day so that students complete academic work using AI-driven applications on screens for roughly two hours each morning, with the remainder of the day devoted to off-screen activities branded as “life skills” development.
In place of teachers, the adults employed by the school are referred to as “guides,” a linguistic shift that mirrors broader efforts within digital education to deprofessionalize teaching and recast pedagogical labor as facilitation rather than instruction. Alpha is rapidly expanding, with new sites operating as private schools as well as cyber charter schools that contract with public districts, thereby drawing on public funds while maintaining private control (Schultz, Reference Schultz2025b).
Alpha School maintains a substantial online presence characterized by polished marketing and a promotional narrative that advances several core claims: that traditional schooling has failed, that AI-based schooling is more fun, efficient, and relevant, and that human teachers represent an obsolete constraint on learning. Promotional videos depict ecstatic children running freely through wooded landscapes, often without visible adult supervision, framing the school as an escape from the disciplinary, bureaucratic, and social constraints of conventional schooling. The school presents itself as an antidote to the alienation associated with public education, suggesting that alienation can be overcome not through richer human relationships or democratic participation but through the substitution of screens and algorithmic systems for teachers and collective pedagogy.
In this vision, freedom is redefined as individualized interaction with technology. Education is stripped of its relational, dialogic, and democratic dimensions and recast as a personalized optimization problem to be managed by AI systems.
In advertisements, media appearances, and on its website, the owner and founder of Alpha School, MacKenzie Price, claims that the proprietary “2 Hour Learning” program produces spectacular gains in academic test scores (Alpha School, 2024a). These bold and endlessly repeated claims are dubious at best. Alpha relies on test score data from its original school in Austin, Texas to assert that its AI-based two-hour academic model yields 2.6 times average learning growth relative to national norms (Alpha School, n.d.-a).
Several problems immediately arise. First, there has been no independent verification of these claims by researchers unaffiliated with the company. Second, Alpha has not released the underlying data necessary for external researchers to independently evaluate or replicate its findings. Third, the company bases its claims on scores from the NWEA MAP assessment, a test designed to measure growth relative to expected national norms (NWEA, 2023). Yet schools vary widely in student demographics, prior achievement, socioeconomic status, and other contextual factors that strongly influence MAP Growth scores. Fourth, MAP norms aggregate performance across schools, districts, states, socioeconomic strata, and racial, ethnic, and linguistic groups, while Alpha’s reported averages are drawn from a very small and socially homogeneous sample of students attending a school overwhelmingly composed of children from high-income families able to afford annual tuition of roughly $40,000.
Despite these limitations, Alpha presents MAP scores from its Austin campus as representative of outcomes across all of its schools and student populations, including significantly lower-cost cyber charter schools and schools serving lower-income communities, such as those in Brownsville, Texas. This claim directly contradicts overwhelming evidence that standardized test outcomes are closely correlated with socioeconomic status. Empirical studies suggest that as much as three-quarters of differences in academic growth scores can be attributed to socioeconomic factors rather than school-level instructional quality (Hu & Morgan, Reference Hu and Morgan2024). Fifth, there are numerous well-established explanations for why students from high-income families consistently outperform their peers that have little to do with the quality of educational programs themselves. These include material conditions such as food security, housing stability, access to healthcare, and exposure to chronic stress, as well as symbolic and cultural factors such as cultural capital and the unequal valuation of knowledge, tastes, and dispositions associated with different class positions (Bourdieu & Passeron, Reference Bourdieu and Passeron1990).
Sixth, and most fundamentally, Alpha’s claims rest on the false assumption that test score growth is equivalent to meaningful learning. Growth scores are representational proxies rather than direct measures of learning. In a school environment where pedagogy and curriculum are heavily organized around constant digital testing through apps, students may become increasingly adept at navigating test formats and optimizing responses. Under such conditions, students may learn how to game assessments without developing deeper understanding, critical thinking, or reflective engagement with subject matter. Test familiarity and score inflation produced through continuous assessment should not be conflated with educational quality or intellectual development.
Alpha Schools has expanded to additional sites in Texas and Florida, with announced or planned schools in Arizona, California, North Carolina, New York, and Virginia (Levy, Reference Levy2025). At the same time, several states, including Pennsylvania, Utah, and Arkansas, have rejected Alpha School’s applications to operate cyber charter schools (Karbal, Reference Karbal2025; Levy, Reference Levy2025).
In Pennsylvania, debates centered primarily on the efficacy of the AI school model. Critics pointed to the poor performance record of cyber charter schools both nationally and within the state when measured by conventional test-based indicators of achievement (Karbal, Reference Karbal2025). Opponents cited a 2019 study of Pennsylvania’s fourteen cyber charter schools, which found that these schools performed worse than or roughly on par with traditional public schools despite higher per-pupil costs (Karbal, Reference Karbal2025). Critics further argued that Alpha’s proprietary two-hour AI instructional model remains largely untested and should not be implemented with public funds without adequate independent study and evaluation.
This focus on efficacy, however, obscures a deeper and more consequential problem: the largely unexamined acceptance, on both sides of the debate, of transmission models of pedagogy and static conceptions of knowledge. By concentrating narrowly on whether AI schools raise or lower test scores, critics implicitly concede the premise that learning is best measured through standardized assessments and that education consists primarily in the efficient delivery of decontextualized and neutral knowledge units. These are precisely the epistemological assumptions that undergird digital educational privatization itself.
What is at stake in the emergence of the AI school is therefore far more than the narrow question of test score performance. Embedded within the model are deeply political and ideological assumptions about the disposable role of teachers, recast as delivery agents rather than socially transformative intellectuals and co-creators of knowledge with students. The model treats knowledge as a deliverable commodity rather than as something produced through dialogue, contestation, and social struggle. It also advances assumptions about the social purpose of schooling as an apolitical instrument of socialization rather than as an institution inherently implicated in struggles over power, meaning, and democracy.
The AI school model further rests on an unexamined faith in the presumed inherent benefits of technology. This faith displaces attention away from the ways digital systems function as instruments of capital accumulation, labor displacement, surveillance, and ideological formation. Taken together, these assumptions reconfigure schooling away from a democratic public good and toward a technocratic apparatus organized around efficiency, compliance, and extraction while obscuring the political choices and interests embedded in the design and use of AI-driven education.
In what follows, I first examine in Section 2.1 how the Alpha School model is explicitly structured to maximize profit, drawing on institutional rationalities and business strategies with clear antecedents in earlier phases of educational privatization. Section 2.2 moves beyond organizational form to interrogate the cultural values, assumptions, and ideologies advanced by Alpha School, including the deployment of corporate culture as pedagogy, the cultural politics underlying claims to political neutrality, and the systematic obscuring of cultural capital in Alpha’s test-based assertions of educational quality. Section 2.3 situates the rise of Alpha and other AI schools within broader economic, political, and cultural transformations, linking the teacherless school to ongoing shifts in labor, governance, authoritarian politics, and capital accumulation.
2.1 The Pursuit of Profit in the AI School
Alpha School’s model rests on the claim that AI technology can teach students better and faster than human teachers, thereby rendering professional educators unnecessary (Alpha School, 2024a). As a for-profit enterprise, Alpha and its related brands are structurally oriented toward minimizing costs and maximizing revenue, a logic familiar from earlier waves of educational privatization.
The extremely high tuition charged at Alpha’s private campuses, approximately $40,000 in Austin, Texas; $50,000 in Santa Barbara, California; $65,000 in New York City; and as much as $75,000 in San Francisco, generates substantial revenue (Alpha School, 2024b). At the same time, Alpha operates a private campus in Brownsville, Texas with tuition reportedly closer to $10,000 and runs cyber charter schools that receive public per-pupil funding comparable to local district allocations (Alpha School, n.d.-b; Schwenk, Reference Schwenk2025). The ability to operate across such radically different price points reveals the scale of profit extraction occurring at Alpha’s elite, high-tuition campuses.
The cyber charter schools operated by Alpha under the name Unbound Academy receive per-pupil funding from public districts that is far lower than Alpha’s private tuition rates and broadly consistent with charter funding levels nationwide (Sitrin, Reference Sitrin2025). District spending on charter schools tends to fall on the lower end of the national per-pupil spending spectrum, in part because high-spending districts rarely contract with charter management organizations that operate on lean budgets and reduced educational services. Professional and ruling-class districts routinely spend far more per pupil than working-class and poor districts. In Fairfield County, Connecticut, for example, the affluent town of Greenwich spends roughly $28,000 per pupil, while the working-class city of Bridgeport spends closer to $18,000 (Connecticut State Department of Education, n.d.). Bridgeport operates six charter schools, while Greenwich has none. Charter schools in Connecticut receive approximately $11,500 per pupil from the state, far below Greenwich’s per-pupil expenditure (Connecticut State Department of Education, n.d.). A district such as Greenwich therefore has little incentive to contract with a charter operator whose business model depends on drastically lower spending.
For-profit charter management organizations thus face strong incentives to minimize expenses in order to maximize profit, and the single largest expense in schooling is labor. Although Alpha Schools does not publicly disclose comprehensive per-pupil expenditure data, it is clear that the company dramatically reduces labor costs by employing no certified teachers, hiring staff without educational credentials, certification, or union representation, and substituting AI systems for professional pedagogical labor. Student-to-guide ratios vary by site, but reported figures underscore the scale of labor reduction. The Alpha School site in Tazewell, Tennessee reports a ratio of approximately twenty-three students per guide, while Unbound Academy enrolls roughly five hundred students per single guide, an arrangement unimaginable in any conventional school model (Sitrin, Reference Sitrin2025).
Texas spends on average approximately $15,000 per pupil in public schools (Texas Education Agency, 2024). In Austin, per-pupil spending is closer to $17,000, while in Brownsville it is approximately $16,000, reflecting a state finance system in which lower-income districts receive a higher proportion of funding from the state to partially offset local revenue disparities (Texas Education Agency, 2024). Teacher labor represents the single largest expenditure for Texas public school districts, accounting for roughly 57 percent of total costs, or about $8,200 of the average $15,000 per pupil (Edison & Reid, Reference Edison and Reid2025). These figures underscore the centrality of professional labor to public education budgets and clarify why labor elimination is such an attractive strategy for profit-seeking education businesses.
Alpha’s labor practices illustrate this logic starkly. The company advertises part-time teaching assistant positions in Austin at an annual salary of approximately $11,250, while simultaneously advertising a “guide” position with a listed salary of $100,000 (Indeed, 2025). This extreme internal wage stratification, combined with the complete elimination of certified teachers, highlights how the AI school model suppresses aggregate labor costs while concentrating compensation in managerial or supervisory roles aligned with corporate control rather than pedagogical expertise. Alpha deepens the tendency of privatized schools to decrease teacher pay while increasing administrator income.
Alpha School has clear historical precedents in earlier waves of educational privatization. Its model closely parallels that of Christopher Whittle’s Avenues: The World School, which pursued a dual strategy of operating elite, high-tuition private schools in major global cities while simultaneously developing lower-cost, scalable charter school replicas with sharply reduced per-pupil spending (Russakoff, Reference Russakoff2015). Whittle, an advertising executive who rose to prominence through Channel One and later Edison Schools, recognized that schools could be branded and scaled like premium consumer goods rather than evaluated primarily on pedagogical substance or educational quality.
Channel One, launched in the late 1980s, famously monetized compulsory schooling by marketing commercial advertising to a captive student audience under the guise of a televised news program (Saltman, Reference Saltman2000). Edison Schools, which later became the largest for-profit charter management organization of the 1990s, similarly pursued scale through branding, standardization, and labor cost reduction (Saltman, Reference Saltman2005; Molnar, Reference Molnar2005). Journalists summarized Whittle’s approach by invoking his own comparison of schools to luxury brands such as Louis Vuitton luggage, emphasizing brand prestige rather than transparency about production processes or educational quality (Russakoff, Reference Russakoff2015).
Alpha’s aggressive pursuit of lower labor costs also echoes Whittle’s early strategies at Edison Schools. Edison initially promoted a progressive-sounding pedagogical model in which students learned by teaching other students, allowing for fewer certified teachers and reduced labor expenditures (Saltman, Reference Saltman2005; Molnar, Reference Molnar2005). This rhetoric was later abandoned in favor of a more rigid and repressive model centered on scripted lessons, standardized curriculum, and test-driven instruction, all designed to produce cost efficiencies through standardization and economies of scale. Despite these efforts, Edison ultimately failed to sustain either educational quality or long-term profitability. Alpha nonetheless reproduces Edison’s early progressive rhetoric in the “life skills” portion of the school day, framing its curriculum as interest-driven while radically reducing time devoted to academic subjects and expanding activities oriented toward entrepreneurship, self-branding, and social-media-based self-promotion. These skills align closely with platform capitalism rather than democratic education.
K12, Inc., which focused almost exclusively on cyber charter schooling, represents another direct precursor to Alpha’s AI school model. K12 sought to replace human teaching with standardized online curriculum delivered at scale, a strategy that produced widely documented low educational quality and bad academic outcomes (CREDO, 2012; Miron & Urschel, Reference Miron and Urschel2012).
Alpha School is distinctive not merely for employing online curriculum, as earlier cyber charter schools did, but for its explicit aim to replace teachers and all academic instruction with AI systems (Alpha School, 2024a). The company operates multiple school brands across private schools, cyber charter schools, and homeschooling platforms, all relying on the same proprietary two-hour AI academic model. NextGen Academy operates as a private school charging approximately $25,000 in annual tuition and combines two hours of AI-based academics with roughly six additional hours of online “gamified” learning and life skills programming (Alpha School, 2024c). These offerings are organized around categories such as teamwork and communication, socialization and public speaking, entrepreneurship and financial literacy, and G.R.I.T., defined as growth mindset, resilience, innovation, and tenacity.
Texas Sports Academy similarly combines the two-hour AI academic block with approximately four additional hours devoted to sports training and life skills, integrating athletic development into the school day while maintaining minimal academic instruction delivered by AI (Alpha School, 2024c). GT School, marketed to “gifted and talented” students, pairs the same academic model with four additional hours devoted to robotics, life skills, chess, and mathematics competition preparation (Alpha School, 2024c). Unbound Academy, Alpha’s cyber charter brand, places students on screens for approximately five of six instructional hours per day, combining online games, roughly two and a half hours of AI instruction, and an additional two hours of project-based activities framed as hands-on life skills (Alpha School, 2024c). Although the program claims that students can pursue interests in STEM, the arts, business, and gaming, these pursuits remain tightly constrained by a standardized AI-driven academic core.
Alpha also operates Novatio School, a fully online program in Arizona with in-state tuition of approximately $7,000, a price point that allows families to cover tuition almost entirely through the state’s Education Savings Account program (Arizona Department of Education, 2024). Finally, Alpha Anywhere functions as the company’s homeschooling brand, extending the same two-hour AI academic model into the household and further blurring the boundaries between public education, private schooling, and privatized domestic labor (Alpha School, 2024c).
The various private, tuition-based Alpha brands are positioned to benefit directly from the expansion of state-based school voucher and Education Savings Account (ESA) programs enacted through Republican-led budgets in 2025. These programs significantly increase the flow of public funds into private, charter, virtual, and homeschooling markets (Lieberman, Reference Lieberman2025; Lieberman, Stanford, & Ifatusin, Reference Lieberman, Stanford and Ifatusin2024). These developments are reinforced by the Trump–Musk–aligned dismantling of the US Department of Education and by proposals to devolve federal education funding to states, where funds can be redirected toward privatization rather than toward compensatory spending designed to address historical and structural inequalities (Office of Management and Budget, 2025).
This shift marks a decisive break with the original intent of the Elementary and Secondary Education Act, which framed federal education spending as a mechanism for mitigating class- and race-based inequities through redistributive support to high-need students and districts. Under current federal education policy, education funds are increasingly steered toward private education businesses, venture-backed AI school models, and voucher-funded tuition programs. In doing so, policy accelerates the transformation of public education from a democratic institution into a privatized investment opportunity.
2.2 The Culture of the AI School: Corporate Culture and Anti-Democracy
Despite its technological novelty, Alpha’s model largely recycles long-standing educational ideologies, repackages familiar tropes, and reproduces problematic trends from earlier phases of education reform. What distinguishes the model is not pedagogical innovation but the fusion of corporate culture with an explicitly anti-democratic orientation. Alpha School exemplifies a form of schooling organized around the values, logics, and hierarchies of the corporation.
Corporate culture is fundamentally incompatible with democratic culture. The business corporation is organized hierarchically to enable owners to extract surplus value from workers, not to cultivate democratic participation, collective deliberation, or shared governance. Even where research demonstrates that more egalitarian and participatory organizations can be more productive, such arrangements remain unattractive to owners because they undermine unilateral control over labor and profits (Bowles & Gintis, Reference Bowles and Gintis2011). Beyond the internal organization of firms, corporate culture also generates and normalizes ideologies that frame schooling as analogous to business and education as preparation for business. Within this logic, students are positioned primarily as future workers, entrepreneurs, and consumers rather than as democratic subjects or critical participants in public life. The expansion of corporate culture into schools therefore occurs at the expense of humanistic, democratic, and egalitarian educational traditions. In these traditions, claims to truth are understood as historically situated, contested, and inseparable from the interests and power relations of those who advance them.
NextGen Academy’s two hours of AI-based academic applications are paired with games for most of the remainder of the school day, effectively replacing education with entertainment and making gamified learning central to the instructional model (Alpha School, 2024c). Gamified learning decontextualizes knowledge from social and cultural context and from student experience, linking learning instead to escapist modes of fantasy often borrowed from mass popular culture, such as arithmetic taught through casting wizard spells at avatars or accumulating points that accrue to be exchanged for digital stickers or other treasures. This approach does not merely borrow the aesthetics of corporate media culture. It imports its ideologies as well, framing learning as consumption, distraction, and extrinsic reward-based compliance rather than interpretation, judgment, or critical understanding (Selwyn, Reference Selwyn2019).
Texas Sports Academy similarly displaces most of the school day with athletics, leaving the two-hour block of online AI instruction as the core of formal academic learning (Alpha School, 2024c). Subjects such as art and music largely disappear, and the model mirrors forms of “unschooling” commonly imposed on child athletes sent to elite training academies. In Open, professional tennis player Andre Agassi described all-day athletic training during childhood as a form of child abuse that foreclosed access to a meaningful education (Agassi, Reference Agassi2009). Although Agassi achieved extraordinary athletic success, he suffered personally, while most children subjected to such regimes lost both their childhoods and their education without attaining elite status. Ironically, Agassi later invested his wealth in charter school real-estate schemes that benefit private investors at public expense (Saltman, Reference Saltman2018).
Alpha’s GT School pairs the two-hour AI academic block with activities heavily oriented toward quantitative subjects, competition, and corporate-style training. The program offers little that cultivates interpretation, judgment, or reflection on self and society. Engagement with the humanities, critical social sciences, and curiosity-driven inquiry is conspicuously absent, foreclosing opportunities for ethical reasoning, historical understanding, and democratic imagination.
Unbound Academy, Novatio School, and Alpha Anywhere operate almost entirely online, largely eliminating sustained human social interaction and stripping away the relational dimensions of schooling. These models remove opportunities for learning rooted in place, community, culture, affect, and embodied experience. They replace these dimensions with standardized, homogenized, and decontextualized forms of instruction optimized for scale, efficiency, and data production rather than meaningful education (Selwyn, Reference Selwyn2019).
Schooling as an expression of corporate culture is evident across both core components of Alpha’s model: the standardized, test-oriented two-hour academic block and the “life skills” and entrepreneurial programming that occupies the remainder of the day. The academic component is framed around several interlocking promises: efficient knowledge delivery through AI rather than teachers; student enjoyment produced by compressing academic labor into a brief daily window; learning purportedly personalized and rendered relevant through algorithmic optimization; and an educational environment that deliberately avoids engagement with political and social issues (Alpha School, 2024a).
Alpha’s founder, MacKenzie Price, has publicly insisted that “we do not let anything political or social come in the way” of classroom instruction (Lombardo, Reference Lombardo2025, n.p.). This posture of purported neutrality has been celebrated by billionaire investor Bill Ackman, who praised Alpha for explicitly rejecting diversity, equity, and inclusion initiatives (Lombardo, Reference Lombardo2025). The issue, however, is not that Alpha’s curriculum and pedagogy are genuinely apolitical. Rather, the values and assumptions embedded in the model align seamlessly with dominant institutions of power, including Silicon Valley technology firms, corporate managerial culture, and the broader MAGA political movement. The claim to neutrality thus functions ideologically. It masks a specific alignment with political-economic interests while disavowing the legitimacy of democratic contestation, pluralism, and critical inquiry within schooling.
Alpha School claims that its two-hour model of academic learning delivered through AI applications is efficient (Alpha School, 2024a). This claim rests on a particular conception of learning in which knowledge is transmitted by technology and consumed by students. It is, in other words, a monological or transmissional model of pedagogy. In this view, knowledge is treated as a fixed object produced by experts and conferred upon students, who are implicitly framed as empty vessels to be filled. Such a conception stands in direct opposition to contemporary theories of culture and education that emphasize constructivist and dialogical understandings of learning (Freire, Reference Freire1972; Giroux, Reference Giroux2011; Hall, Reference Hall1997).
Where transmissional models treat knowledge as static, constructivist approaches understand knowledge as actively produced through dialogue between teachers and students and situated within specific cultural, historical, and social contexts (Freire, Reference Freire1972; Giroux, Reference Giroux2011; Hall, Reference Hall1997). Transmissional models nonetheless remain attractive within systems of educational governance because they lend themselves to quantification. When knowledge is framed as something that can be delivered, consumed, and displayed back to authorities, it can be readily measured through standardized testing.
Indeed, the standards and accountability movement of recent decades has normalized standardized testing as the primary indicator of educational quality. Educational technology has intensified this tendency by increasingly making testing itself the mode of pedagogy. So-called personalized learning platforms continually assess students as they move through content, embedding testing within the learning process and producing constant quantified representations of progress (Roberts-Mahoney, Means, & Garrison, Reference Roberts-Mahoney, Means and Garrison2016; Selwyn, Reference Selwyn2019). This arrangement enables not only measurement but also extensive tracking, monitoring, and comparison of students over time.
The monological view of pedagogy thus aligns seamlessly with technology-based instruction. Dialogical forms of pedagogy, by contrast, require human interaction, interpretation, and relational judgment. Dialogue-based teaching allows educators to respond to students in real time and to relate knowledge to particular cultural contexts, subjectivities, and lived experiences. A fundamental limitation of AI-based pedagogy is that, despite persistent rhetoric of personalization, AI systems cannot meaningfully relate objects of study to student experience or to the broader social, political, and cultural forces that shape that experience.
Alpha also promises student enjoyment and motivation through the radical reduction of academic learning time, implying that traditional academic study is inherently disengaging (Alpha School, 2024a). Across its marketing materials and promotional media, the company repeatedly suggests that academic subjects should be minimized and compressed, while AI-mediated learning is framed as more pleasurable, efficient, and engaging than conventional schooling. Central to this claim is Alpha’s heavy reliance on gamified learning, including at least one school model organized almost entirely around video gaming (Alpha School, 2024a, 2024c).
English Language Arts within Alpha’s model relies heavily, and in some cases exclusively, on AI-driven reading platforms such as Lexia, a program originally designed as a supplementary intervention rather than a standalone curriculum (Lexia Learning, n.d.). Language instruction is also delivered through applications such as Duolingo, which employ point-based competition, gamification, and animated characters to incentivize engagement (Duolingo, 2024).
Far from apolitical, the organization of the school day reproduces a commercial and anti-intellectual cultural logic in which sustained academic inquiry is displaced by instrumental activities tied to work, productivity, optimization, and consumption. The bulk of the day is organized around programs that resemble corporate training modules, which Alpha labels “life skills.” These include time-efficiency techniques such as the Pomodoro method, which breaks students’ labor into optimized bursts of productivity, alongside wellness programming that emphasizes individual self-regulation in response to social conditions beyond individual control (Alpha School, 2024a). The company frames this extensive life skills programming as focusing not on the accumulation of knowledge but on using it (Raviglia, Reference Ravaglia2025). Given the sharply limited space for academic study or dialogic exchange, however, it remains unclear what substantive knowledge students are expected to use, or toward what collective or democratic ends.
Corporate language and assumptions dominate the remainder of the school day. As one local report notes, “the rest of the school day is spent developing life skills like public speaking, leadership, teamwork, and entrepreneurship through workshops” (Fox 7 Austin, 2024), along with instruction in financial literacy (Lanum, Reference Lanum2025) and activities in which students design startup pitches (Raviglia, Reference Ravaglia2025). Alpha circulates a steady stream of promotional materials reinforcing this corporate vision of schooling. A Miami press release describes a school day in which “8-year-olds launch startups, 10-year-olds give TED-style talks and 12-year-olds tackle Harvard Business School challenges,” beginning with a Tony Robbins–inspired event called “Morning Launch,” where students engage in goal setting and group motivational exercises (Community News Releases, 2024a).
According to Alpha School Miami head of school Alex Kiser, a former Google employee, “Our students love the workshops and real-world challenges we put in front of them. We’re preparing them for jobs that might not even exist yet” (Community News Releases, 2024b). What goes unaddressed in this celebration of workforce preparation is the widespread displacement of future jobs, including teaching itself, by the very AI technologies around which Alpha is organized. Also absent is any engagement with the democratic question of how people in a technologically mediated society will collectively determine which forms of labor are socially necessary, valuable, or desirable.
Many of the technology magnates who finance, promote, and ideologically defend these systems, including Peter Thiel, Elon Musk, Marc Andreessen, Jeff Bezos, Balaji Srinivasan, and Alex Karp, have openly sought to weaken democratic governance in favor of technoauthoritarian forms of rule (Slobodian, Reference Slobodian2023). This vision prioritizes technological acceleration over democratic control of social priorities. Alpha’s curriculum appears deliberately structured to avoid the kinds of learning that would cultivate the knowledge, dispositions, and critical capacities required for collective deliberation over production, consumption, and public life.
The so-called foundational skills emphasized in Alpha’s model are not rooted in traditions of scholarship, sustained study of history and society, or critical interpretation of texts. Nor do they involve dialogic engagement with teachers and peers around questions of meaning, ethics, or social responsibility. Instead, the emphasis on corporate culture provides little preparation for democratic social relationships, collective self-governance, or reflective participation in public life. Alpha’s model does not merely presume that learning and democracy are unrelated. It actively reconceptualizes schooling as an institution subordinated to the production of a future corporate labor force.
This orientation is reinforced by Alpha’s explicit embrace of online consumer culture, particularly the practices of self-presentation, branding, and performative labor normalized through social media platforms. A job advertisement for Alpha’s Austin campus illustrates how fully this logic has been absorbed into the school’s pedagogical model (Crossover, 2025). The advertised position, with a salary of approximately $100,000, does not seek an educator or a specialist in any academic discipline. Instead, Alpha seeks a “social media expert” with a significant online presence who will teach students how to “build powerful personal brands.”
As discussed above, Alpha does not hire subject-matter teachers because academic expertise is presumed to reside exclusively in AI applications. What it does hire are individuals who, in the company’s own words, “live and breathe social media strategy, content creation, and community building,” and who are tasked with working directly with children. This raises a critical question: What kind of community is being cultivated, and toward what ends?
As one job posting states:
You’ll work directly with students to help them build engaged audiences using cutting-edge AI tools and proven growth tactics. Think of yourself as a blend between a TikTok strategist, content coach, and startup advisor, guiding 50 emerging personal brands simultaneously.
Learning in Alpha Schools thus appears to involve training students not only to become permanent self-promoters or “brands” but also to function as marketers for the school itself. Students spend substantial portions of the school day producing content that the company uses to enhance Alpha’s social media visibility and recruit additional student-customers. Children are reimagined as workers engaged in unpaid promotional labor for a for-profit education company. One point not to be missed is that academic content is presumed to originate entirely from digital applications, while human labor within the school is redirected toward branding, audience cultivation, and platform growth.
The knowledge and skills valorized in this job advertisement affirm and intensify the entertainment-driven and self-promotional logic of social media culture. Empirical research and critical social theory have repeatedly shown that pervasive social media use among youth is associated with heightened screen dependency, anxiety, depression, attentional disorders, and affective dysregulation (Berardi, Reference Berardi2017; Crary, Reference Crary2022; Saltman, Reference Saltman2025b). In response to these concerns, schools across the United States have increasingly moved to ban or sharply restrict mobile phone use during the school day to mitigate excessive screen exposure and its documented harms (Pew Research Center, 2024).
Alpha School, by contrast, does not merely tolerate screen saturation. It celebrates and institutionalizes it, making social media logics central to pedagogy, curriculum, and the formation of subjectivity itself. School culture is organized around platforms, metrics, and visibility, displacing human interaction and democratic values with the dominant ethos of Silicon Valley: machines over people, profit over human well-being, technological acceleration over ecological sustainability, and the synthetic over the natural world. What is particularly notable here is that Alpha’s model includes no intellectual tools to criticize, analyze, or interpret the values, ideologies, assumptions, and interests in social media or promoted by social media forms of communication. It is strictly affirmational. The lack in the curriculum of serious engagement with the traditions of the humanities, social sciences, critical media literacy, or critical pedagogy result in students who have no space within the model to question, debate, or discuss how such technologies and their content produce them as particular kinds of subjects, mobilizing particular identifications and desires.
Alpha School consistently denigrates academic learning as “boring” and portrays traditional schooling as fundamentally broken. In a promotional video on the company’s website, founder and owner MacKenzie Price asserts: “Traditional school is broken. It’s outdated, full of busywork, and sadly for our kids, often a waste of time” (Karbal, Reference Karbal2025). Alpha and its media boosters further claim that the model is uniquely “personalized” and “relevant.” A Forbes article that reads largely as promotional content states: “By allowing students to work individually at their own pace but within a shared environment, Alpha School preserves the social benefits of schooling while eliminating many of its traditional inefficiencies” (Raviglia, Reference Ravaglia2025). Appearing on Fox News, Price similarly claimed: “We use an AI tutor and adaptive apps to provide a completely personalized learning experience for all of our students” (Lanum, Reference Lanum2025). A Miami press release reposted by Alpha asserts that “the entire school experience is shaped around individual student needs, aptitudes, and passions” (Community News Releases, 2024a).
In this context, “personalized learning” functions as a misnomer. While AI applications allow students to move more quickly or slowly through standardized curricular sequences, such pacing adjustments do not constitute personalization in any meaningful educational sense. Moving at one’s own pace simply means consuming pre-designed curriculum at different speeds. What is systematically excluded from this model is the relational production of knowledge through interaction between teachers and students and the shaping of learning through cultural specificity, lived experience, and social context. The applications do not enable learning that is meaningfully connected to students’ lives, communities, or historical conditions.
Instead, these systems presume that the knowledge to be delivered is neutral, universally valuable, and beyond contestation, removed from cultural politics altogether. In practice, the technology is structurally incapable of the dialogic engagement through which students learn to interpret claims to truth in relation to what is meaningful and consequential in their lives beyond school. What the applications do provide is incessant assessment, as testing becomes the mode of pedagogy itself. The logic of the standards and accountability movement, with its false claims to objectivity and neutrality, is repackaged as personalization and attentiveness to student subjectivity (Williamson, Reference Williamson2017).
Alpha’s claim to personalization thus participates in a broader corporate school reform pattern that appropriates the language of progressive education while evacuating its substantive commitments (Saltman, Reference Saltman2023). The applications and programs employed by Alpha are not “relevant” in the progressive educational sense of relating learning to students’ lived experiences, cultures, and social realities. Instead, relevance is narrowly redefined as entertainment, gamification, and distraction, reducing learning to pleasurable activity rather than meaningful understanding. By contrast, progressive and critical education traditions ground learning in students’ experiences while situating those experiences within broader social, political, cultural, and economic contexts. These traditions begin with what students know and recognize, but they do not end there. Knowledge becomes a means through which students reinterpret their experiences and understand how those experiences are shaped by larger social forces and structures. Learning, in this view, cultivates agency and the capacity to use knowledge to act upon and transform the social world. Knowledge becomes meaningful insofar as it is socially critical and potentially transformative (Giroux, Reference Giroux2011).
Alpha’s corporate culture training and AI-driven curriculum advance a very different conception of relevance, personalization, and engagement. Here, relevance is equated with alignment to commercial interests, corporate work cultures, and ideologies of self-optimization for productivity. Absent is any sustained relationship between learning and self-understanding, learning and social understanding, or knowledge as a basis for collective action to address public problems. In short, knowledge for agency is displaced.
Alpha’s model not only evacuates the social and political significance of learning; it promotes a radically individualistic and consumerist conception of education. Knowledge is framed as something to be consumed, displayed, and leveraged for academic advancement and eventual labor market insertion. One can only hope that the forms of labor students are encouraged to desire do not include teaching itself, since the Alpha model systematically renders teachers obsolete.
In interviews across television and digital media, MacKenzie Price repeatedly describes Alpha School as politically neutral, claiming that it deliberately avoids political and social issues in the classroom, which she portrays as a central failing of traditional public schools (Lombardo, Reference Lombardo2025). The centrality of technology to Alpha’s model allows its promoters to draw upon the long ideological legacies of positivism and techno-utopianism. Positivism frames truth as a collection of supposedly objective facts while erasing the values, interests, and social positions that determine which facts count, which facts matter, and how facts are interpreted. Within this framework, only that which can be quantified, measured, and controlled is recognized as legitimate knowledge.
Positivism has played a central role not only in the rise of the standards and accountability movement over the past four decades but also in the contemporary celebration of educational technology. AI-based and other digital learning platforms readily lend themselves to standardization, homogenization, quantification, and a false appearance of disinterested objectivity (Selwyn, Reference Selwyn2019). When student activity is rendered measurable by an application, the resulting numerical outputs are presented as neutral indicators of learning. In doing so, they conceal the values, assumptions, ideologies, and social positions embedded in both the curriculum and the technologies that deliver it. AI thus adds an additional layer of false neutrality on top of the already deceptive neutrality claimed by standardized testing regimens.
When Alpha relies on platforms such as IXL to teach core subjects, students are given no opportunity to challenge, debate, engage in dialogue, or reject the range of acceptable truth claims embedded in a lesson (IXL Learning, n.d.). An eleventh-grade social studies lesson, for example, frames economic systems as limited to two options: market economies, misrepresented as driven exclusively by business and consumer decisions, and command economies, in which governments dictate all economic activity. Workers, collective decision-making, and democratic economic alternatives are absent altogether. A student familiar with the work of University of Massachusetts economist Richard D. Wolff might recognize that actually existing democratic economic formations, including worker-owned cooperatives in places such as Silicon Valley, directly contradict this narrow framing. Yet the standardized structure of IXL provides no mechanism to introduce, debate, or even acknowledge such perspectives. The curriculum designers remain invisible, and there is no pedagogical space for contestation.
As instruction shifts further toward AI-mediated systems, this problem intensifies. There is often no human interlocutor available to question at all. The limitations of such technologies include epistemic boundedness, in which answers are restricted to singular or narrowly defined responses that foreclose alternative reasoning; the absence of sources or arguments justifying claims; and the concealment of the decision-making processes that shape how “truth” is presented. As with standardized testing, knowledge appears to come from nowhere, treating truth as an effect of authority rather than interpretation and evidenced argument. These pedagogical conditions also shape student dispositions. Emphases on efficiency, completion, and surveillance encourage compliance over understanding and discourage questioning for fear of penalty.
These systems normalize dominant assumptions and ideologies as beyond dispute while denying students the means to interrogate them. As AI platforms scrape the web for information, they frequently prioritize speed and responsiveness over accuracy, producing confident but erroneous outputs. In early 2025, for example, when I asked ChatGPT whether any schools were taught primarily by AI, it incorrectly reported that no such schools existed even though there was an abundance and variety of material available about Alpha from academic journals, the popular press, and on the web. As for-profit products, these platforms are designed to satisfy institutional consumers such as school owners and administrators who prioritize quantified outputs and performance metrics over dialogue, exploration, or meaningful learning.
As Alpha misrepresents its model as neutral, objective, and outside social and political concerns with justice, it becomes essential to ask what kind of social and educational vision the school advances and whose interests that vision ultimately serves.
2.3 Situating Alpha School in Relation to Technoauthoritarianism and Digital Educational Privatization
Alpha School must be understood as more than a technological novelty or an educational trend. It is better situated within the broader political, economic, and cultural forces shaping contemporary capitalism. Alpha functions both as an expression of and a contributor to the consolidation of technoauthoritarianism. Leading figures associated with this formation, including Peter Thiel, Elon Musk, Marc Andreessen, Balaji Srinivasan, and others, openly reject democratic accountability and advance visions of platform sovereignty, corporate rule, and market governance insulated from democratic control. As Quinn Slobodian documents in Crack-Up Capitalism, these actors increasingly imagine capitalism without democracy, treating democratic governance not as a safeguard but as an obstacle to accumulation (Slobodian, Reference Slobodian2023).
Within this worldview, society is imagined as a collection of radically self-reliant individuals stripped of public goods, collective provision, and democratic self-governance. Social life is reorganized around privatization and de-democratization. For technoauthoritarians, technological acceleration, often framed as “effective accelerationism” and justified through speculative benefits for hypothetical future populations, takes precedence over human needs, democratic values, and ecological survival in the present. As Naomi Klein and Astra Taylor (Reference Klein and Taylor2025) and Slobodian (Reference Slobodian2023) argue, this vision is organized around the principle of exit. Tech elites seek to withdraw from democratic society by creating privatized zones such as charter cities, gated enclaves, offshore havens, underground bunkers, and even space colonies, insulated from taxation, regulation, and democratic oversight. It’s worth noting, as Slobodian (Reference Slobodian2023) points out, that the vision of charter cities and exit was inspired by charter schooling and the dream of carving out privatized entities from the public.
Technoauthoritarians routinely naturalize inequality, framing it as biologically determined or justified through discredited notions of intelligence and cognitive hierarchy. The public is increasingly portrayed as irrational, dangerous, or in need of control. Accordingly, technoauthoritarians invest heavily in surveillance, policing, and military technologies designed to manage and repress populations rather than enable democratic participation. This orientation is visible in the work of Palantir, the data analytics and AI firm led by Alex Karp, which develops surveillance and military systems for state and corporate clients (Palantir Technologies, n.d.). At the level of policy, similar logics are evident in initiatives such as Elon Musk’s DOGE project, which sought to dismantle and privatize core functions of the US federal government, replacing public provision with market-driven alternatives (Klein & Taylor, Reference Klein and Taylor2025). These projects simultaneously generate new profit opportunities while eroding democratic authority, transparency, and accountability.
Alpha School aligns closely with this technoauthoritarian imaginary. As Katya Schwenk observes in Jacobin, Alpha resembles a “Silicon Valley fever dream,” populated by promotional scenes in which “a ten-year-old boasts that he is a successful Airbnb manager, while another constructs a miniature Cybertruck. High schoolers take ‘business calls,’ lounging in a WeWork-style open office space” (Schwenk, Reference Schwenk2025, 1). As Schwenk further notes, Alpha has received support from tech elites, including SpaceX, which has provided programming and transportation assistance to an Alpha-affiliated school near its South Texas campus (Schwenk, Reference Schwenk2025, 2).
Alpha shares with technoauthoritarian ideology a commitment to private-sector provision, withdrawal from the public sphere, and the abandonment of civic responsibility. As Klein and Taylor describe, advocates of “exit” argue that the ultra-wealthy should be freed from obligations to the public altogether, including taxation, regulation, and democratic constraint: “Retooling and rebranding the old ambitions and privileges of empires, they dream of splintering governments and carving up the world into hyper-capitalist, democracy-free havens under the sole control of the supremely wealthy, protected by private mercenaries, serviced by AI robots, and financed by cryptocurrencies” (Klein & Taylor, Reference Klein and Taylor2025, 1).
In the case of Alpha School, exit takes the form of rejecting public education and its democratic mission. Rather than preparing students for collective self-governance, public deliberation, or shared responsibility, Alpha models schooling on the corporate workplace. Its curriculum privileges entrepreneurial subjectivity, self-optimization, and market rationality while marginalizing academic traditions that cultivate historical understanding, critical reflection, and democratic imagination. Alpha thus functions as a training ground for a radically individualized and asocial future, one in which government is displaced by AI, public institutions are supplanted by private platforms, and education serves capital rather than democracy.
Alpha also participates directly in the political project of the contemporary right by advancing voucher-based privatization, economic exclusivity, and an educational system in which familial wealth increasingly determines access to quality and opportunity. In this sense, Alpha School does not merely reflect technoauthoritarianism. It actively helps to reproduce the cultural dispositions, labor relations, and political subjectivities required for its expansion.
The preceding analysis demonstrates that AI schooling, as exemplified by Alpha School, cannot be understood as a misguided reform or a neutral technological experiment. It represents a convergence of digital educational privatization, child labor extraction, corporate culture as pedagogy, and a broader technoauthoritarian project that seeks to displace democratic education with marketized, depoliticized, and extractive forms of schooling. If AI schools clarify what is at stake in the unchecked expansion of educational technology under contemporary capitalism, they also clarify what must be resisted and reimagined.
The final section therefore turns from critique to possibility. It asks how AI might be disentangled from labor exploitation, privatization, and authoritarian epistemologies and instead reoriented toward critical, democratic, and non-extractive educational practices. This reorientation cannot be achieved through technical fixes alone. It must be understood as a political, pedagogical, and cultural project rooted in democratic values, collective agency, and the refusal of child labor in digital form.
3 From Labor Extraction to Critical AI Education Practice
The prior sections have shown that dominant AI education products extract child labor by conscripting children into the work of becoming data-generating engines. They have also shown that the emerging trend of AI schools realizes a long-standing project of automating teacher labor in the service of educational privatization, profit-seeking, and the expansion of corporate culture and neoliberal ideology into schooling. Widely used AI education platforms largely reproduce uncritical and anti-democratic transmissional models of pedagogy, denying cultural politics, power relations, and material struggle as constitutive dimensions of knowledge and learning.
This section argues that although most AI education initiatives currently foster authoritarian social relations, institutionalize transmissional pedagogy, and erode the democratic capacities of public schools, AI technologies are not inherently destined for these ends. Under specific and tightly circumscribed conditions, they can be repurposed as tools for critical pedagogy. Examples of genuinely critical AI education remain rare. Accordingly, this section first maps the dominant terrain of AI education platforms to clarify their authoritarian and anti-critical tendencies. It then examines what is presently being promoted as “critical AI education,” identifying both its possibilities and its limits. Finally, it turns to forms of cultural production, particularly technology-infused art and activist projects, to illuminate alternative ways AI can be mobilized to expose power rather than conceal it.
This analysis is undertaken in the hope of advancing critical and democratic uses of technology in schools capable of challenging technoauthoritarianism and neoliberal fascism (Giroux, Reference Giroux2019) while cultivating democratic dispositions and social relations across educational and civic institutions. Section 3.1 examines how widely adopted AI education programs conceal power, foreclose inquiry into the relationship between knowledge and domination, and sever learning from student experience, social context, material interests, and broader structures of inequality. In their dominant forms, these programs foster dispositions aligned with authoritarian epistemologies while simultaneously extracting commercially valuable data through uncompensated child labor and appropriating teacher work. Section 3.2 turns to a range of AI pedagogy projects that, despite being marketed as “critical,” lack core elements of critical pedagogy and critical theory. The section then examines the work of the art and activist collective Forensic Architecture, which utilizes AI to make visible the otherwise obscured operations of power, violence, and state and corporate wrongdoing. These projects move beyond the transmissional AI pedagogy model characteristic of dominant digital education infrastructure and platforms while offering varying degrees of democratic pedagogy, critical agency, and capacity for transformative public intervention.
3.1 Dominant AI Programs
3.1.1 AI Education as Standards and Accountability Inertia
Teacher use of AI is expanding rapidly in public schools. As of 2023, AI education programs were used by approximately 18 to 25 percent of K–12 teachers (Slama et al., Reference Slama, Pane, Steiner, Hamilton and Pane2019). By the 2024–25 academic year, roughly 60 percent of teachers reported using generative AI for lesson planning, rubrics, quizzes, and professional communication such as emails (Gallup & Walton Family Foundation, 2025). AI education tools can be broadly grouped into two categories: educational technology infrastructure that predates generative AI and generative AI applications that operate through and reinforce this existing infrastructure.
In mapping dominant AI and AI-adjacent technologies in schools, my purpose is to demonstrate that these tools overwhelmingly extend rather than disrupt the logic of the standards and accountability movement that preceded them. That movement sought to reduce what counts as teachable to what can be measured, to recast teachers as delivery technicians, and to treat knowledge as a commodity that can be efficiently transmitted. Drawing on positivist ideology (Adorno, Reference Adorno1999) and the dogmas of scientific management (Giroux, Reference Giroux1983), standards and accountability reforms systematically delinked knowledge from the conditions of its production and from the social, political, and power relations that shape it.
As a result of these positivist foundations, the standards and accountability movement denied the theoretical bases of both practice and fact. It rejected the reflective praxis through which teachers relate knowledge to students’ experiences and social contexts to make learning meaningful, socially critical, and potentially transformative (Giroux, Reference Giroux2011). Such praxis is a necessary condition for education to become a basis for social and political agency, enabling learners to interpret public problems, translate private troubles into collective concerns, and act together to address them (Mills, Reference Mills1959; Bauman, Reference Bauman1999).
Contemporary digital education technologies largely reproduce this legacy by advancing claims of neutrality, universality, and disinterested objectivity about knowledge and human interests. These framings obscure central questions: Who produces this knowledge? Whose interests does it serve? What assumptions and ideological commitments shape its construction? Despite the rhetoric of “personalized learning,” adaptive digital platforms rarely relate knowledge to student subjectivity or social context. Instead, they transmit standardized curriculum while adjusting only the pace or difficulty of delivery. These systems also embed continuous standardized assessment into the pedagogical process itself. This is not incidental. As detailed below, most generative AI platforms in education are deployed through, and constrained by, these adaptive, test-centered infrastructures.
The standards and accountability movement accelerated sharply between 2000 and 2010 with the passage of No Child Left Behind and the subsequent adoption of the Common Core State Standards. During this period, testing and textbook corporations spent millions lobbying for laws requiring curricular alignment with test-based standards, policies that secured billions in profits for education conglomerates such as Pearson NCS, Houghton Mifflin, McGraw-Hill, and ETS (Strauss, Reference Strauss2015).
Just as momentum behind ever-expanding standardized testing appeared to be weakening amid the growth of the opt-out movement (Hursh et al., Reference Hursh, Deutermann, Rudley, Chen and McGinnis2020) and student walkouts protesting high-stakes testing, digital applications and so-called personalized learning platforms began entering classrooms at scale. Despite their reformist rhetoric, these programs functioned primarily as delivery systems for canned, homogenized curriculum taught through near-constant testing.
Under the banner of efficiency, innovation, and personalization, the core logics of the standards and accountability movement were not displaced but intensified. Educational technologies deepened the reduction of learning to measurable outputs while embedding continuous testing into everyday pedagogy. In doing so, they established the infrastructural foundation upon which contemporary AI education programs would later be built. These systems do not break from standards and accountability reforms but rather extend and automate them.
Infrastructural educational technologies are adaptive and datafied systems that function simultaneously as assessment and instructional platforms. They translate student and teacher activity into data and normalize learning through algorithmic processes. Often marketed as “adaptive” or “personalized,” these platforms continuously measure, sort, and compare students. Some of the most prominent examples include the following.
i-Ready, produced by Curriculum Associates, combines diagnostic assessment with instructional content and is used by approximately 14 million students, or roughly one-third of K–8 public school students in the United States. As a privatized benchmark for student “readiness,” i-Ready displaces teacher judgment and professional autonomy, transferring evaluation of student capacity to proprietary algorithms owned by a for-profit corporation.
IXL Learning reported usage by approximately 17 million students as of 2025 (IXL Learning, 2025). The platform extracts extensive student clickstream data. Although marketed as a supplemental learning tool, IXL frequently functions as de facto curriculum, as illustrated by its wholesale adoption by Alpha Schools and many public school systems. Despite being described as adaptive, the platform delivers a standardized and homogenized curriculum, adjusting only the pace of delivery while remaining unable to relate knowledge to cultural context, student experience, or socially meaningful purposes.
Renaissance Star Assessments and Freckle, used in over 32,000 schools nationwide, integrate student placement, performance measurement, and adaptive curriculum into a single system. In these platforms, standardized assessment becomes the mode of pedagogy itself. Quantified measures of learning are presented as neutral indicators of knowledge acquisition, obscuring the cultural politics of curriculum and the ideological assumptions embedded in test design.
NWEA MAP Growth, used in approximately 20,000 K–8 schools, uses predictive analytics to algorithmically sort and track students. The platform conceals the politics of knowledge by comparing student performance to statistical norms rather than engaging students in dialogue, debate, or questioning of truth claims. Knowledge is framed as something to be absorbed and returned correctly, with deviations from the norm triggering interventions designed to improve test performance. These systems do not merely deny the relationship between knowledge and power. They actively undermine critical intellectual dispositions such as curiosity, judgment, dissent, and creativity, capacities that are also foundational to democratic participation.
Achieve3000 Literacy, produced by McGraw-Hill, is a “differentiated” reading platform incorporating generative AI elements. It exemplifies how legacy educational publishing firms central to No Child Left Behind have integrated AI to maintain market dominance. In practice, differentiation largely means accelerating or decelerating a prefabricated reading sequence rather than engaging student subjectivity or social context in ways that make reading meaningful or transformative.
DreamBox Learning, now part of Discovery Education, is a widely adopted adaptive mathematics platform with millions of users. Like other integrated assessment-instruction systems, DreamBox combines standardization with transmissional pedagogy. Mathematical knowledge is decontextualized and treated as a set of discrete skills rather than as a tool for understanding public problems or lived social conditions. Critical mathematics pedagogy, by contrast, situates math as a resource for collective agency and social analysis. Adaptive math platforms largely foreclose these possibilities.
Amplify is a for-profit literacy program used across numerous states. Its algorithmic scoring and instructional guidance are inseparable from its curriculum design, resulting in the displacement of teacher curricular autonomy by standardized, pre-packaged content delivered through transmissional pedagogical models.
Taken together, these integrated assessment-instruction platforms hardwire for-profit technology agendas, data extraction, standardization, and banking education assumptions into the daily routines of schooling. In doing so, they undermine teacher autonomy over pedagogy and curriculum while constraining the capacity to teach for critical, democratic, and public purposes. These platforms are typically adopted and rendered mandatory by administrators influenced by technology marketing and techno-utopian ideology (Means, Reference Means2018). Teachers are required to subordinate professional judgment and critical reflexivity to corporate systems while simultaneously being placed under continuous technological surveillance as they measure students.
Informed by positivist ideology (Giroux, Reference Giroux1983/2025; Saltman & Means, Reference Saltman, Means, Waite and Bogotch2017), the metrics generated by these platforms present themselves as neutral, disinterested, and universally valuable. In reality, they conceal the human decision-making, values, and ideological commitments embedded in their design and algorithms. The key point not to be missed is that the newer generation of generative AI education programs is largely layered onto this existing data-driven infrastructure. It constitutes the dominant landscape of contemporary AI education, not a rupture from it.
3.1.2 Generative AI Programs Plugged into the Digital Infrastructure
Most generative AI programs currently entering schools are not freestanding pedagogical innovations. They are designed to plug into, and operate through, the pre-existing digital infrastructure of standards, accountability, and datafication. The following examples illustrate how generative AI is largely layered onto this infrastructure rather than disrupting it.
Khan Academy and Khanmigo
As of 2025, Khan Academy products were used in approximately 550 US school districts (Richmond Forum, n.d.). Khanmigo, a GPT-based AI tutor, reported roughly 700,000 student and teacher users during the 2024–25 academic year (K–12 Dive; Education Week). Crucially, Khanmigo is integrated with NWEA MAP Growth through “Learning Paths,” binding it directly to the standards and accountability testing infrastructure.
Although Khan Academy maintains a nonprofit status, it has long been supported by venture philanthropy (Saltman, Reference Saltman2010, Reference Saltman2023), including funding from the Gates Foundation, venture capitalist John Doerr through the New Schools Venture Fund, and Google. Khanmigo itself operates through for-profit GPT architecture and is integrated with the for-profit NWEA platform. This configuration exemplifies philanthrocapitalism’s fusion of nonprofit and for-profit educational entities (Saltman, Reference Saltman2018). By yoking AI tutoring to MAP Growth, Khanmigo anchors AI instruction within a framework defined by neoliberal instrumentalism, transmissional pedagogy, metrics-driven accountability, and the denial of the cultural politics of knowledge. Learning is severed from student subjectivity, social context, and collective agency.
MagicSchool AI
MagicSchool AI claims approximately six million teacher users across more than 20,000 schools, including state and regional contracts (MagicSchool AI, 2025). The platform offers AI-generated tutoring, lesson plans, grading tools, and administrative supports, positioning itself as a centralized AI hub for curriculum production and instructional management. In doing so, it promotes teacher deskilling and dependency on AI-mediated pedagogy.
In my own testing, I used MagicSchool’s chatbot to develop a middle-school social studies lesson on Guatemala in the early 1980s, including US support for the genocidal Ríos Montt regime and the political-economic interests involved. I found that the AI chatbot required numerous refinements, extensive prompting, and a great deal of prior knowledge on the part of the user to avoid having the algorithm whitewash the mass murder, torture, rape, and disappearances of predominantly indigenous Ixchil and Mayan people in the highlands. The system encouraged alignment of the lesson with state standards, and then readily removed discussion of racism and imperialism to comply with US state gag laws. When prompted to generate a Marxist lesson without standards alignment, it complied. When prompted to generate a fascist lesson, it initially refused, then produced the content after the request was reframed comparatively.
What is notable here is the degree of expertise, time, and intervention required of a teacher to prevent historical whitewashing, as well as the political contingency and inconsistency of the system’s responses. Teachers using these tools have no visibility of, or control over, the ideological filters, safety constraints, and political assumptions embedded in the platform. Nor do they have any way of knowing how these constraints may shift over time in response to political pressure or market incentives.
MagicSchool’s student-facing math tutor performed substantially worse. In a ninth-grade polynomial lesson drawn from Chicago Public Schools curriculum, the system became incoherent, truncated explanations mid-response, substituted letters for numerical exponents, and ultimately failed as an instructional tool. Across uses, the platform proved unstable and unreliable, frequently freezing or misprocessing inputs.
SchoolAI
SchoolAI markets itself as a FERPA- and COPPA-compliant “safe” AI environment built around closed or “walled-garden” chats, dashboards, and AI spaces. It is primarily marketed to districts seeking risk management rather than pedagogy. Like other so-called personalized learning platforms, SchoolAI adapts pace rather than meaning. It cannot relate knowledge to student experience, culture, or social context. More significantly, it represents an extreme case of vendor-governed education, in which ideological filters, epistemic limits, and curricular boundaries are embedded in opaque code inaccessible to teachers, students, or administrators.
Publisher-Embedded Generative AI Suites
Major educational publishers, including McGraw-Hill, HMH, Savvas, and others, have embedded generative AI tutors, automated content generation, and adaptive analytics into their existing digital curricula. McGraw-Hill advertises generative AI add-ons across its K–12 platforms, claiming reach in the millions. These integrations consolidate publisher market dominance while locking AI use into standards-aligned, required curriculum. Teachers are further subordinated to accountability infrastructures, and the possibility of deploying AI for critical pedagogy is structurally foreclosed. The radical potential of generative AI to situate knowledge historically, politically, ethically, and economically is neutralized by design.
Creatium (Formerly Prof Jim)
Houston Independent School District uses the for-profit AI platform Creatium to generate large volumes of curriculum materials, including reading passages, video lessons, and AI coaching tools (Hernandez, Reference Hernandez2024). This represents a clear case of a major US district outsourcing curriculum production to a private AI vendor. Students are positioned as data engines, performing the new child labor through behavioral tracking. Teacher labor is displaced as interaction, assessment, and coaching are replaced with analytics dashboards and subscription services. Teachers are repositioned as content feeders, while their labor is repackaged and sold back to districts. Student behavioral data becomes the core resource justifying subscription value, deepening technological dependency and privatizing pedagogy.
Ambient and Unofficial For-Profit Generative AI in Schools
General-purpose language learning models such as ChatGPT, Gemini, and Copilot are widely used by teachers, with approximately 60 percent reporting AI use for professional tasks (Gallup & Walton Family Foundation, 2025). These tools operate largely outside formal procurement processes, constituting a gray-market form of digital educational privatization in which the cognitive labor of teaching is informally outsourced to corporate AI systems. While less bounded than district-approved platforms, these tools still displace teacher judgment, cannot relate learning to lived student experience, and risk intensifying technological dependency.
Generative AI can assist teachers in situating knowledge within broader political, historical, and economic contexts. It cannot, however, relate knowledge to student subjectivity or lived experience. Critical pedagogy is dialectical, requiring mediation between subjective experience and objective social structures. AI can only support such work when guided by teachers with well-developed critical capacities and substantive knowledge, as illustrated in the Guatemala lesson example. Moreover, while current safety frameworks in some platforms prohibit overt dehumanization or historical falsification, these safeguards are mutable and vulnerable to political pressure, capital interests, and authoritarian capture. As AI dependency deepens within educational institutions, the risk of democratic loss of control over knowledge production intensifies.
3.2 Critical AI Pedagogy
A number of AI pedagogy initiatives describe themselves as “critical.” At the outset, it is essential to clarify that the term critical carries two very different meanings in educational discourse. In its most common usage, “critical thinking” refers to generic problem-solving or analytical skills. Used in this way, the term adds little to the concept of thinking itself. By contrast, critical theory and critical pedagogy name a distinct set of intellectual and political traditions that explicitly relate knowledge and culture to power, politics, ethics, and history.
Critical theoretical traditions, including the Frankfurt School, feminism, socialism, radical democracy, critical philosophies of race, and political economy, situate knowing within class antagonisms, cultural struggles, and material and symbolic interests. These traditions link subjectivity to social structure and treat knowledge as historically produced and politically contested. Accordingly, critical pedagogies cultivate practices of interpretation and meaning-making that enable students to understand how claims to truth are implicated in relations of authority and domination, as well as in struggles for resistance and emancipation. They foreground how particular interpretations of experience and social reality express the interests and perspectives of some groups rather than others.
From this standpoint, critical pedagogy is necessarily suspicious of positivist and radically empiricist truth claims that deny the interests and ideological positions of those producing knowledge or that present facts as theoretically innocent. In the spirit of critical theory, critical pedagogy engages in sustained critique of truth claims by situating them within broader political, economic, and cultural forces, structures, and tendencies. It also encourages students to theorize experience itself. Subjective experience, from this perspective, is not transparent or self-authenticating but requires interpretation of how subjects have been socially formed to make meaning. By situating truth claims historically and politically, learners can apprehend the interests and ideologies that shape meaning-making practices. Ultimately, critical pedagogy aims to expand egalitarian and just social relations.
Critical pedagogy is an intersubjective, embodied, and dialogic practice oriented toward expanding collective human agency through deepened self-understanding and social analysis. AI platforms are structurally incapable of engaging in critical pedagogy. They are disembodied, lack consciousness and agency, cannot participate in reflective praxis, and cannot engage students as co-subjects in meaning-making. Moreover, critical pedagogy seeks collective liberation through the transformation of consciousness as a basis for emancipatory social projects. AI systems possess no commitments beyond those programmed into them, have no material engagement with social life, and cannot participate in collective struggle to transform institutions or societies. Rather than emancipatory, AI platforms operate through logics aligned with neoliberal governance and authoritarian forms of educational rationality.
3.2.1 “Critical” AI Education Programs
Several AI pedagogy initiatives claim to foster critical engagement through the use of AI. Closer examination, however, reveals that most employ an impoverished and depoliticized conception of criticality.
AI Pedagogy Project (Harvard)
The AI Pedagogy Project provides instructors with resources, sample assignments, and prompts intended to promote “critical engagement” with AI, including its social implications, capacities, and limits. Yet the project defines criticality in narrowly liberal terms. Its framing emphasizes “diverse perspectives” and the avoidance of “political bias” while acknowledging that technologies are not value neutral.Footnote 3
What is conspicuously absent from this formulation is sustained engagement with the relationship between knowledge and power or with the cultural politics of knowledge within and beyond technology use. The project does not begin from the premise that knowledge and technology are expressions of particular values, assumptions, and ideologies, nor that these symbolic formations are tied to material interests, institutional power, and struggles over domination and resistance.
This omission is consequential. Competing interpretations are not simply different perspectives; they are rooted in distinct social positions, interests, institutions, and relations of power. A review of sample assignments on the project’s website reveals no sustained engagement with these dynamics. Instead, assignments range from liberal pedagogical exercises, such as conducting Socratic dialogues with AI or asking AI to speak from the standpoint of fictional characters, to neoliberal tasks such as business analyses of corporate profit maximization or engagement with World Economic Forum reports on the “future of work.”
References to critical scholarship, critical traditions, or foundational critical concepts are notably absent. Despite this, ChatGPT itself described the AI Pedagogy Project as critical in the following terms: “… this project invites educators and students to ask who built the AI, what values are embedded, how it might shape teaching and learning, and whose interests it serves, as well as how students might become agents rather than passive recipients.” While it is possible that instructors already grounded in critical theory could appropriate these materials for more critical ends, the project itself does not facilitate such work. Nor does it connect interrogation of AI to broader political-economic structures, cultural struggles, or material antagonisms. This absence fundamentally limits its critical capacity.
Navigating AI (FLOE – Flexible Learning for Open Education)
The Navigating AI project offers resources focused on inclusive education, emphasizing open-access digital technologies and AI literacies that help educators and students understand how AI systems generate outputs and how they are structured. The project aims to foster informed and responsible AI use and to promote student agency.
Funded by the William and Flora Hewlett Foundation, the initiative is broadly affirmational toward the expansion of digital technologies. Many of its supported projects emphasize inclusion of marginalized populations, particularly people with disabilities, within AI systems. Some materials raise important questions about power, including who controls AI, whose knowledge is embedded, how open or closed AI ecosystems are, and whether AI adoption should be informed rather than uncritical. Recurring themes include accountability, transparency, privacy, information literacy, equity, largely understood as preventing intensified sorting and sifting, and ecological sustainability.
What remains largely absent, however, is guidance on using AI to engage cultural politics in educational practice or to situate AI use within broader political-economic struggles and structures of domination. The project does not move from critical literacy about AI toward collective intervention in public problems or democratic struggle. Notably, when offering examples of “critical AI,” Navigating AI points to the AI Pedagogy Project at Harvard, thereby reproducing the same limitations.
A project that does not suffer from these constraints is Forensic Architecture, to which I now turn.
3.2.2 Critical Education AI Projects
Forensic Architecture is a multidisciplinary research, cultural, artistic, and political activist platform that conducts investigations using digital technologies, including machine learning. Its projects examine state violence and the corporate profiteering intertwined with it, exposing the often-obscured operations of power by tracing relationships among governments, corporations, and civilian populations. Investigations frequently involve reconstructing scenes of violence using evidence drawn from witness testimony, digital traces, devices, online platforms, and recovered physical objects.
One of Forensic Architecture’s most widely recognized projects, Triple Chaser, investigated the global deployment of tear gas manufactured by the Safariland corporation and used against civilian populations by police and military forces, including in occupied Palestine and along the US–Mexico border. Witnesses documented spent tear gas canisters with cell phones, and machine learning systems were trained to identify these canisters, allowing researchers to scan the internet for additional instances of their use. The project revealed the global scale of deployment and mapped the connections between Safariland’s owner, Warren Kanders, and the governments purchasing the weapons. Triple Chaser was exhibited at the Whitney Biennial in New York City, where Kanders served as chair of the board. The findings were amplified by the activist group Decolonize This Place, and sustained public pressure ultimately led to Kanders’s resignation.
Other Forensic Architecture projects include reconstructions of mass killings in the Sinai carried out by the Egyptian state, investigations into police killings of civilians, journalists, and political figures, analyses of dimensions of the genocide in Gaza, and studies of repression targeting higher education protesters globally, particularly those mobilizing against the genocide in Gaza.
Several features distinguish Forensic Architecture as a model for the critical use of AI and digital technologies in educational contexts. Its projects investigate violence and injustice perpetrated by powerful institutions, recovering suppressed knowledge while producing new public knowledge. These investigations serve the public interest rather than private, commercial, or captured state interests. Crucially, Forensic Architecture relates objects of analysis, such as tear gas canisters, not only to the lived experiences of those subjected to violence but also to the broader systems, institutions, and political-economic structures responsible for its production and deployment. Knowledge production is thus transformed into a vehicle for social and political agency.
Forensic Architecture frequently collaborates with Bellingcat, an open-source investigative journalism collective. Together, these projects embody an epistemological and pedagogical orientation fundamentally opposed to that of the standards and accountability regime and its contemporary manifestations in AI education platforms. Whereas dominant AI education treats knowledge as something to be transmitted, measured, and optimized, Forensic Architecture and Bellingcat treat knowledge as something to be discovered, constructed, contested, and mobilized. Where AI education platforms position technology as the authoritative source of knowledge and teachers as facilitators of delivery, these projects position participants as both learners and producers of knowledge, with technology functioning as an investigative instrument rather than an epistemic authority.
Dominant AI education platforms are structurally incapable of engaging learner subjectivity or social and cultural context, thereby de-linking social forces from the people most affected by them. By contrast, Forensic Architecture explicitly links the subjectivity and social positioning of those harmed by violence to the institutions, actors, and systems that produce it. Technology is thereby repurposed as an instrument not only of inquiry and understanding but of collective political intervention. This stands in sharp contrast to prevailing AI education models, which commodify knowledge as discrete units to be delivered, assessed, and reproduced through constant testing, reducing learning to credentialing and narrow economic reward.
Forensic Architecture and Bellingcat should therefore be understood as exemplars of critical AI practice. They center public problems, injustice, and inequality; deploy technology in the service of democratic inquiry and collective agency; and affirm the capacity of people to investigate, interpret, and produce knowledge that intervenes in the social world. These projects reject transmissional pedagogy that renders learners passive recipients. Instead, participants become social and political actors.
Democratic societies require citizens capable of critically interpreting knowledge and transforming it into collective understanding and action. Education has always been foundational to democracy. As AI use in education expands rapidly, its potential for critical and democratic practice remains real but contingent. Whether that potential is realized depends on educators and cultural workers grounding AI use in critical theory and pedagogy. Claims to truth must be interpreted in relation to the interests, ideologies, and social locations of those who advance them. Knowledge must be understood as dynamic, contested, and co-constructed through dialogic exchange among unequally positioned actors. It must be engaged in relation to the broader social forces that give it meaning, something AI can sometimes help illuminate, and to the lived experiences and cultures of students, something AI itself cannot do.
4 Conclusion: Contesting the New Child Labor
This Element has argued that the expansion of AI education platforms and AI schools marks the emergence of a new form of child labor. Through the compulsory use of digital platforms and AI systems, students are increasingly conscripted into the production of commercially valuable data that sustains profit-driven educational infrastructures. This labor is unpaid, largely invisible, and naturalized through the language of personalization, efficiency, and innovation. Unlike earlier forms of school commercialism, digital educational privatization extracts value directly from students’ present activity rather than merely shaping future consumer or worker dispositions. Children are no longer only being prepared for work. They are already working.
This new child labor must be understood in relation to broader political-economic transformations. It is embedded in global commodity chains that link mineral extraction and super-exploited labor in the Global South to data extraction and cognitive labor in schools in the Global North. It is also aligned with the rise of technoauthoritarianism, a political project that seeks to displace democratic governance with privatized technological control, to hollow out public institutions, and to normalize inequality as efficient and inevitable. AI schools and dominant AI education platforms do not simply adopt new technologies. They reorganize schooling around automation, surveillance, and extraction, extending transmissional pedagogies and positivist epistemologies that deny the cultural politics of knowledge.
At the level of pedagogy, AI education platforms intensify the legacy of standards and accountability reforms by embedding continuous assessment, surveillance, and data production into everyday learning. These systems undermine teacher autonomy, deskill educators, and erode the democratic capacities of schools by severing knowledge from student experience, social context, and collective agency.
Yet technology is not destiny. While AI systems are structurally incapable of doing critical pedagogy themselves, this book has shown that digital technologies can be repurposed within critical educational practice. Projects such as Forensic Architecture demonstrate that AI and machine learning can be used to expose power, recover suppressed knowledge, and transform investigation into collective political agency. These projects function as counter-models to extractive and authoritarian AI education, illustrating what it means to place technology in the service of public problems rather than private profit.
The future of AI in education will not be determined by technology alone, but by political struggle. Whether schools become data factories for authoritarian capitalism or spaces for critical and democratic learning depends on the actions of teachers, students, unions, communities, and publics willing to contest educational privatization and reclaim schooling as a democratic public good. In many cases such an aspiration must involve refusal of bad educational products and coordinated efforts to stymie and shut down the digital education profiteers through activism, policy, legislation, and popular educational projects. However, this struggle also requires teachers to be understood not merely as users or interpreters or critics of technology but as cultural workers capable of participating in the design and production of better alternative technological projects. Working collaboratively with students, artists, activists, and technologists, educators can help develop investigative, dialogic, and socially engaged digital practices that foster critical pedagogy rather than extraction.
Recognizing the new child labor is therefore a necessary political and educational act. Naming it disrupts the common-sense narratives that frame AI education as neutral, inevitable, or emancipatory. The task ahead is not to humanize extractive technologies but to challenge the conditions that make extraction appear natural. This means insisting that technology, in its educational uses and beyond, serve radically democratic self-governance, egalitarian social transformation that links multiple struggles for justice, and collective freedom rather than the freedom of capital.
Kenneth J. Saltman
University of Illinois Chicago
Kenneth J. Saltman is Professor of Educational Policy Studies at University of Illinois Chicago. His recent books include Smart Drugs, Attention Doping, and Screen Addicts: The Drug Attention Industrial Complex in Education (Bloomsbury 2025); The Politics of Education: A Critical Introduction, 3rd Edition (Routledge 2025); The Corporatization of Education: Selected Writings of Kenneth J. Saltman (Routledge 2024); The Disaster of Resilience: Education, Digital Privatization, and Profiteering (Bloomsbury 2023); The Alienation of Fact: Digital Educational Privatization, AI, and the False Promise of Bodies and Numbers (MIT Press 2022).
Alexander J. Means
University of Hawaiʻi at Mānoa
Alexander J. Means is Associate Professor and former Chair of the Department of Educational Foundations at the University of Hawaiʻi at Mānoa. His interdisciplinary research draws on critical traditions in the social sciences and humanities and focuses on the intersections of educational policy, political economy, technology, and the cultural politics of education, with particular attention to how education relates to questions of power, inequality, and social transformation. He is currently the Editor-in-Chief of the Review of Education, Pedagogy, and Cultural Studies, an international peer-reviewed journal published by Taylor & Francis. His most recent book is Teaching Against the Machine: Critical Pedagogy in an Era of Artificial Intelligence (Bloomsbury 2026).
About the Series
Elements in the Cultural Politics of Education is dedicated to critical analysis of cultural and political issues in education. This collection features concise, accessible, and intellectually rigorous books that intervene in some of the most pressing debates and dilemmas shaping contemporary education. Each volume offers an incisive examination of a distinct issue or problem in education be it pedagogical, institutional, cultural, ideological, or political.
The series is grounded in a recognition that education is never neutral and that it represents a key site where subjectivities, values, knowledge, and futures are being shaped and struggled over.
Volumes in the series are politically engaged and practically attuned while also theoretically ambitious. They are written for scholars, students, educators, and engaged publics concerned with the future of education in times of crisis and conflict while remaining focused on specific problems ranging from privatization and authoritarianism to the climate crisis. Ultimately, this series insists on the relevance of education to struggles for social justice, collective agency, and radical democracy in the twenty-first century.
