The Odd Omission of the Arts from the Digital Reorganization of Society

Amid our growing anxiety about the contributions of digital technology to the corrosion of truth, democracy, and mental well-being, relatively less attention has been paid to how internet-enabled computing has influenced what is often termed the “fine arts,” as well as the performing arts. While the initial rise of the commercial internet, its successor the social web, and the emerging decentralized web (or Web3) have all explicitly addressed questions of culture and entertainment, “the arts” as a conceptually coherent but broadly conceived professional and institutional practice of aesthetic production has not substantially figured into the conversation.

Among contemporary critics of various aspects of the way in which the internet has impacted society, the digitization of the arts—and especially the performing arts—has hardly received a passing glance. Often the worst that is said about digital performing arts content is that it’s aesthetically wanting. The topic rarely arouses interrogation, much less indignation. At the same time, the way in which the Covid-19 pandemic induced many performing artists and performing arts organizations to utilize technology to retain a connection with audiences was often cast as the embrace, whole cloth, of the language of techno-utopianism and the triumph of techno-determinism. Without any sense of self-awareness or even unease, many performing arts institutions proudly proclaimed the power of digital technology to reach new audiences and expand access to cultural offerings. ^{Footnote 1}

In this context, two ideas are worthy and urgent objects of discussion within the arts and among the collection of social critics, scholars, and activists who have forcefully questioned the rampage of digital technology into every facet of our social, economic, and political lives.

The first idea is that during a time of profound self-reflection in the arts, the assimilation of digital computing technology into the production, distribution, and consumption of human creative expression raises profound political and ethical questions about the status of artists, systems of aesthetic production, and the social role of arts and cultural institutions. Our contention is that the arts need to regard digital computing technology not merely in terms of audience and revenue access, but also as raising foundational political questions about the purpose and quality of artistic production.

The second idea is that the digitization of creative production reveals essential features of the overall digitally driven social reorganization currently underway. That is, to the extent we are seeking to grapple with how digitization is reshaping our society, its impact on the arts can help to illuminate the mechanisms of this transformation—as well as the stakes.

Our argument is two-fold. First: transformations to human creative expression wrought by commercially driven digital systems both clarify and presage an ascendant logic that prioritizes generating an increasing share of monetizable human attention as both the precondition and the goal of all digitally mediated activity. That is, this imperative is not only the objective of digital activity; it also shapes the productive impulse from which such activity springs. Second: this logic implants vulnerabilities in other systems of social value. This capitalist logic, left unchecked and unquestioned, has the potential to degrade or even wipe out the institutional systems designed to realize and protect other social goods. In the context of the arts, this logic operates as a kind of bait and switch. Promising the development of new audiences, broader accessibility to artworks, and new revenue streams for artists, digital platforms may actually produce many of the same damaging outcomes we have seen in other contexts: the reduction of everything into “content,” further concentration of corporate control over creative expression, the erosion of authoritative human discernment, and the intensification of human labor.

In making this argument, we address a wide variety of cultural expressions online, from the digitally microscopic act of “liking” someone else’s video or image to a classically trained ballerina posting a video of herself dancing to Beyoncé; from a theatre company live-streaming Shakespeare to a visual artist striving to amass a fan base on Instagram. These forms of expression are neither aesthetically equivalent nor comparable in magnitude and mode, and the phenomena we describe apply more to some of them than to others.

Not only do the digital production, distribution, and consumption of these forms of creative expression share a unifying digital capitalist logic, but that logic, in fact, depends on erasing all distinctions in content, whether it is considered high (fine) or low (entertainment) art, professional or amateur.

From Copyright to Commodification

In the late 1990s and early 2000s, much of the commentary and scholarship regarding digitally enabled creative production focused on questions of copyright and ownership (see Lee et al. Reference Lee, Laidlaw and Mac Sithigh2015). Digital creative content distribution arose alongside the commercial internet. Sometimes this distribution concerned materials openly available for public use, but the most notorious examples—think of music-sharing services, most famously Napster—trafficked in creative works in which someone or some entity could assert ownership. Major corporate content publishers, particularly TV and movie studios and record labels, were especially alarmed, fearing that paying audiences would find the allure of internet “piracy” hard to resist.

The principal result of these concerns was a series of international copyright treaties that were domestically ratified through the Digital Millennium Copyright Act of 1998 (DMCA). The law provides the major contours of how content is managed online. Among its chief provisions:

(1) rightsholders can set the terms of access; (2) any violation of access restrictions, implemented through electronic means, is criminally liable, whether or not the copyright itself has been infringed upon; (3) providers of digital services can avoid liability for copyright, so long as they take down offending content if notified by rightsholders.

This regime has been scrutinized and criticized on a number of fronts. A key issue among scholars has been the way in which the DMCA furthers a long-term evolution of US copyright doctrine to privilege content rightsholders over content users—and to ensure that commercial publishers, not individual artists, remain the dominant rightsholders. Siva Vaidhyanathan, for example, has remarked that, for most of the last century, “courts and corporations have exploited public concern for rewarding established authors by steadily limiting the rights of readers, consumers, and emerging artists,” explaining that “[a]ll along, the author was deployed as a straw man in the debate” (2001:25). On Vaidhyanathan’s account, by granting rights-holders the ability to set the terms of digital access, the DMCA favors corporate rights-holders, while eviscerating the “public bargain between producers and users” that enables creative production to serve a larger public good—one that relies upon actual use subsequent to the act of authorship or production (222–23).

Since the DMCA, the debate about the role of copyright in enabling or inhibiting the creative possibilities latent in digital technology has been extensive and engaged (see Vaidhyanathan Reference Vaidhyanathan2001; see also Lessig Reference Lessig2008). In their 2022 polemic, Chokepoint Capitalism, legal scholar Rebecca Giblin and longtime commentator on digital trends Cory Doctorow blame the DMCA copyright regime for creating digital “walled gardens” that, they argue, disproportionately benefit a few publishers and service providers while reducing choices, freedom, and (digital) mobility for the rest of us. According to Giblin and Doctorow, the DMCA facilitated the growth of digital rights management (DRM) regimes through such applications as Apple Music and Amazon Kindle, systems that lock users into using Apple or Amazon hardware, respectively, to access their content (2022:25–27).

Unlike many contemporary critics who focus on the ways technology companies manipulate and sell human attention through social engagement and data extraction, Giblin and Doctorow do not see this as the central problem with digital technology. They argue instead that “Big Tech abuses monopoly power to deprive us of choice by limiting what we can buy, redirecting our searches to hide rivals’ products, and locking us into its ecosystem with technologies we can’t alter without a lengthy prison sentence” (2022:36).

For a time, the oddly inverse relationship between legal enforcement and the actual prevalence of unauthorized content distribution and use (“online piracy”) seemed to undermine Giblin and Doctorow’s concerns. Napster, which launched in 1999 and has been long seen as the avatar of the illegal content sharing era, was shut down in 2001—the same year that Apple debuted iTunes, its DMCA-compliant music distribution system. As soon as Napster shuttered, however, new file sharing tools immediately appeared, such as LimeWire and BitTorrent. Other copyright circumvention tools like “cyberlockers” have also emerged. ^{Footnote 2} Like much illegal activity, assessing the prevalence of online piracy is difficult. Legal scholars João Pedro Quintais and Joost Poort suggest that, whether looking at actual piracy data or the impact on content sales (in music especially), it is clear that rates of piracy only continued to grow in the early 2000s—well after the adoption of the DMCA regime (Quintais and Poort Reference Quintais and Poort2019:812–13).

In an important shift that ultimately seems to confirm Doctorow and Giblin’s fears, Quintais and Poort argue that the more recent real decline in piracy is due to the widespread availability of cloud applications such as Netflix and Spotify that provide ubiquitous access to content—both streaming and downloading—at affordable rates. In terms Giblin and Doctorow might embrace, what Quintais and Poort seem to intimate is that, if the price of entering the walled gardens is low enough, most of us will be more than happy to entrap ourselves within them.

This brief history illustrates how digital technology companies actively manage the interplay between technological and legal tools to strengthen their control over creative expression. While many of these same systems of control had previously been exerted by analog gatekeepers like TV stations and music publishers, what matters now is how a distributed, many-to-many information network like the internet changes the modality—and therefore the reach—of control over expression.

But what these accounts do not explain is the continued development of the digital creative commodity form.

This is where we pick up the story.

Creativity on an Industrial Scale

The most vivid illustration of how creative expression is being subsumed into the commercial internet has been the recent and stratospheric emergence of the so-called creator economy.

A digital creator is anyone who originates content online: whether a blog post or a comment on a blog, a picture or a video. The subject matter could be anything. The format or platform is irrelevant. The intent to make money from the content does not disqualify anyone from consideration as a “creator.” Yet to say that a creator is merely someone who does something online is a semiotic maneuver that misses a subtle but significant difference between a “creator” and an “influencer.” In seeking to distinguish between a “creator” and an “influencer,” technology reporter Kyle Chayka defines the creator as someone who produces the raw material of a content-driven internet:

Though not mutually exclusive categories (a creator can also be an influencer and vice versa), a creator’s intentions are oriented toward an act of digital production, rather than toward self-promotion. While both may pull the same economic levers—the monetization of human attention through advertising—the purpose and form of their digital participation sets them apart.

The venture firm SignalFire estimates that approximately 50 million people are deriving some income from the productive act of creating for the internet. Of these, about 2 million are making content full-time (Yuan and Constine n.d.). Venture funding for the services that enable the creator economy jumped to over $1 billion in 2021 (CB Insights Reference Insights2021).

All artists utilizing online platforms to produce or promote their work are creators, though clearly not all digital creators are artists. When venture capitalists talk about the creator economy, the arts may be the last thing on their minds. They are often referring to the digital hobbyist, like a home cooking instructor, a rare lizard enthusiast, an amateur makeup artist, or someone inviting you to GRWM (“get ready with me”) on TikTok. But the subject position of the digital creator illuminates a logic that we predict will relentlessly dominate efforts to colonize the totality of human creative expression online (and, perhaps, offline as well).

To bring a commodity of the creator economy into being and enable its exchange, digital firms have brought to life a new kind of commodity fetishism. The “secret” of commodity fetishism, first described by Karl Marx in Capital ([1867] 1990), is the process through which the social relations of production become submerged in a tradable good that presents its value as intrinsic rather than as the product of human labor. This is an important concept for Marx, because it accounts for the process divorcing value from labor, allowing for the aggregation of surplus value by the owners of capital (164–65). The sale of bottled water vividly illustrates this concept. Water is a freely available resource, supplied in most communities as a public good whose costs of capture and distribution are shared by residents through local taxation. Bottling water—which can then be marketed as “spring” or “purified” or “enhanced” water, therefore better for us than tap water—and selling it through a retailer obscures the natural availability of water and converts water into a product marketed interchangeably alongside other finished comestible and potable goods. What is important to recognize is not that some public goods are vulnerable to privatization, but rather that we come to understand water as something that is exchanged in a marketplace, rather than as a freely available community resource. Instead of seeing water as a distinct substance whose presumptive context is either free or shared use, we come to regard water as the same kind of thing as soda or juice—or even as chips or bread.

In the digital creator economy, commodity fetishism takes on a different character. The shorthand is “user-generated content,” but this term obscures a novel interplay between labor and capital.

The first aspect of digital commodity fetishism is the commodification of expression. This stage comprises using digital media services—YouTube, Instagram, TikTok—to convert expression, either planned or serendipitous, into digital content. All of these services are equipped with ever-evolving editing features such as filters, cropping, and tools to add text, music, or animation that make it radically easy to capture what you do (or say, or perform), modify your output to make it more entertaining to a potential audience, and then publish it for all to see and hear. As a human social behavior, this process is quotidian, even mundane. You can’t walk down the street without passing someone taking a selfie or video, presumably for a social media platform.

But, as a matter of political economy, this process is a radical act. Once the image, video, text, or sound has been uploaded and digitally modified, it has been transformed from a rich social context into a digital object that—at the layer of code—is functionally indistinguishable from the billions of other pieces of “content” that populate a platform. This content can then be seamlessly accessed across the platform, from a feed or a search engine, and compared objectively to other kinds of

content. As Kate Eichhorn writes:

With the click of a button, this process converts an act of creative expression performed by a human being into a digital “thing” connected to a user. The productive context has been effaced. And, at this stage, the digital thing—content—can be connected seamlessly from one user, the content producer, to many other users, the content consumers. Critically, as Eichhorn says, this process makes all kinds of content the same. All other qualia—intention, context, affect, substance—are erased. Every bit of content is just another video, image, or post.

If the story ended here, it would describe what the technology companies hail as “disintermediation”: removing a gatekeeper—a newspaper editor, a cultural institution, a publishing house—from the exchange between a producer and consumer, enabling one-to-many and many-to-many relationships of production and consumption. At the surface, this describes nearly every social media platform, especially the vast ocean of people uploading pictures of themselves and commenting on pictures of others.

But the story doesn’t end here.

The second aspect of digital commodity fetishism is the monetization of creative expression, in which the algorithmically generated encounter between the digital content and an audience (between user as producer and user as consumer) is abstracted. In its conventional form, commodity fetishism abstracts the social relations of production through the seemingly objective artifact of the market price. In digitally enabled creation, the social relations of production are reduced to a share of attention: the size of an audience, measurable levels of “engagement” (number of and duration of views, likes, and so on), and the composition of the audience.

The abstraction of audience to attention share is a double movement.

First, it’s an abstraction in and of itself, as the share of attention represents a value that can be directly exchanged with an advertiser for money. Sometimes this happens through digital marketing agencies and “influencer” marketplaces, but often digital platforms manage this process directly. ^{Footnote 3} The underlying digital content is irrelevant to this exchange.

For the exchange to work, advertisers must be able to trust that any and all content alongside or through which their brand appears also aligns with their economic and ethical sensibilities. To achieve this, many technology firms employ machine learning and artificial intelligence agents alongside the digital economy equivalent of child laborers toiling on the shop floor. In this case, the degraded and dehumanized laborers are content moderators, who are often employed as independent contractors and who may be in developing countries, well out of sight. This global underclass wades through the ethical excrement of humanity, often at the expense of their own mental well-being, ensuring that images of child sexual abuse or abject violence do not appear in our content feeds (Chotiner Reference Chotiner2019). ^{Footnote 4} By filtering out attributes that would make the social context of the content impossible to ignore, the final stages of computational and human filtering remove frictions that might remind us that the world is not exclusively made up of photogenic meals and vivacious young people looking hot at the beach. As Kate Klonick reminds us, “[t]hough corporate responsibility is a noble aim, the primary reason companies take down obscene and violent material is the threat that allowing such material poses to potential profits based in advertising revenue” (2018:1562).

Functionally, this filtering renders all digital content truly interchangeable from the perspective of the advertiser seeking to acquire human attention. Indeed, the advertiser’s real interest is the audience that is represented by the content: how big is it, who are they, how much money can they spend, and how engaged are they? Often, this movement is regarded as essentially indistinguishable except for scale and speed from the now century-old sale of attention through mass media—radio, television, and print. Yet this comparison misses the way in which the democratization of digital content introduces the very problem that content moderators are hired to solve. ^{Footnote 5}

Also distinct in a digital context is a second movement, the re-abstraction of its viewers into sets of data integrated with billions of other data points about who they are and how they behave under certain conditions. The inferences gleaned from the computational analysis of that data are in turn used to refine how engagement is generated and sold, either through innovation in the services offered to users or in the services that facilitate advertising. Kate Eichhorn explains: “consider what happens when you’re asked to upload an image of yourself to a digital platform. On its own, your photograph isn’t valuable. If a platform has millions of tagged photographs just like yours, however, the photographs can be mined and, in turn, used to drive research and development” (2022:17). As Julie Cohen further elucidates, “[c]ultivated and extracted data enter an industrial production process during which they are refined to generate data doubles—information templates for generating patterns and predictions that can be used to optimize both online and physical environments around desired patterns of attention and behavior” (2019:64).

In the classical model of industrial capitalism, the fetishism of commodities enables the aggregation of surplus value by capital owners as they pocket the difference between the commodity’s price in the market and the total cost of labor. In the digital creator economy, the creator becomes an industry unto themself through a final abstraction: the industrialization of creative expression. This third aspect of digital commodity fetishism is enabled through patronage (crowdfunding, subscriptions, tips, etc.) and sales channels (courses, merchandise, newsletters, direct engagement, etc.). In effect, these digital services intensify the act of digital commodity creation, absorbing greater time and effort from the creator and therefore subsuming more and more of the creator’s labor power into exchangeable digital objects. (So, for example, a home chef posting instructional videos may send a newsletter with recipes, may sell or resell cooking equipment, or may offer “master classes” for paid subscribers.)

Crucially, this digital industrial logic does not inhere within an institutional apparatus separate and apart from the conventional structures of sociality (“private life”). Contrast the 21st-century digital creator economy to legacy creative industries, such as the Hollywood studio system or recording labels. In those arrangements, devised in the 20th century, the extraction of surplus value in creative industries happened exclusively within a formal industrial system. In the 21st century, this model of industrial control over creative expression has swelled both in its reach and in its productive capacities. Today, anyone with a smartphone is a potential producer; and, any producer equipped with a digital device is an industrial production system unto themself.

A tragic example of this phenomenon is Ryan Kaji, a child YouTube star who turned unboxing products into a brand. Born in 2011, Kaji has been making “toy videos” since he was three. (The videos are a kid explaining his new toys.) As of 2023, “Ryan’s World” had 32 million subscribers and Kaji’s net worth was $35 million, deriving from an endless stream of now professionally produced videos across a diverse set of YouTube channels (Kang Reference Kang2022). Even though Kaji’s performance is the product, it is clear that he, too, is an instrument of a corporate enterprise. (“During the shoot in Simi Valley, after a long stretch of filming in the intense sun, I overheard a crew member say to him, ‘If you finish this scene, you can play Minecraft’” [Kang Reference Kang2022].)

While the digital platform exerts absolute control over the commodification, monetization, and industrialization of creative expression, it does not pay for the production itself. Instead, the platform bears the cost of developing and maintaining the tools and protocols that enable creators to voluntarily produce their own content. In exchange, the platform takes a percentage of what revenue-generating content earns in the attention marketplace.

The commodification, monetization, and industrialization of creative expression produce an abstraction of creative production three times over. First, the act of production is abstracted into a digital object produced by a user. When a person uploads a video of themselves lip-synching a song to TikTok, that video turns into a digital object that, from the perspective of TikTok, is completely interchangeable with any other video. Second, the social consumption of the content by other users (now users-as-consumers) is abstracted into an “audience” that is traded in an open market as a share of attention. In this market, content is compared not according to its rich social context, but instead according to metrics assigned by the platform—often focused on vectors of what the industry characterizes as engagement, such as popularity (“likes”), stickiness (how long on average users-as-consumers viewed the content), or recency (how recently on average users-as-consumers interacted with the content).

Third, the attention share of these users is then subjected to computational organization through increasingly vast data engines, whose output is used to refine techniques to further manipulate user attention towards still greater engagement. The digital platform is ultimately seeking to generate an understanding of “people like you” in order to increase the frequency and amplitude of your views and interactions with content. In a digital context, this process is exponentially more granular than forms of market research conducted by older media (Netflix sorts its audience into more than 1,300 “taste communities” [Rodriguez Reference Rodriguez2017]) and can adjust path dependencies at the individual user level instantaneously.

The result of these three processes is that the content-producing individual themself—the user-as-producer—is abstracted into a commercial enterprise. The user-as-consumer is no longer digitally interacting and transacting with a real human being, but instead with a range of products and services that, collectively, may represent the output of aggregated labor power. Yet, crucially, unlike the traditional corporate form, the enterprise presents itself as an unmediated human-to-human relationship. We—all of us—become the embodiment of the rationality of the corporation.

The critical question, then, is whether and how this compounded abstraction influences creative expression, either in the creative process itself or in the experience of the audience.

What’s New about Industrial-Scale Digital Creativity?

The capitalist logic of attention capture is neither generally new nor especially new to the arts. The phenomenon of mass media is well-studied. As audiences expanded by means of radio, film, and television, new opportunities arose for monetizing attention (e.g., “ratings”). These developments have transformed the arts since the early 20th century. The performing arts, in particular, have been variously produced, distributed, and consumed through mass media—often fomenting great debate about the aesthetic consequences.

Beginning with the earliest oral traditions and evolving through the establishment of formal venues and the professionalization of theatre, music, and dance, the performing arts have long pioneered converting attention into profit, whether monetary or as some other functional good, such as reputation, social cohesion, or political persuasion.

As with many digital phenomena, the question is whether the scale, instantaneity, and ubiquity of the internet intrinsically influence quality or whether they merely alter quality through quantity. To put it another way, are things online different by virtue of something about digital technology itself, or just because digital technology enables so much more of everything?

We believe there are five features specific to digitally mediated creative expression that deserve independent analysis. (We postpone for the moment the question of whether the results of these features are desirable.)

The Missing Politics of Digital Creativity

Over the past decade, arts communities and institutions have increasingly been engaged in a period of intense self-examination over issues of cultural representation, labor conditions, and environmental impact. Institutions especially are grappling with a long overdue appraisal of the ethical implications of the arts they sponsor. The critiques are both inward-facing and outward-facing, touching their internal operations (workplace culture, staffing and hiring, pay, working hours, and more) and their programming, communications, and outreach. Leading institutions have begun to grapple with the political reputations and global impact of their donors, the racial composition of their leadership, or their approach to programmatic decisions.

Individual artists, too, have participated in this fieldwide interrogation. Some appear as “newly admitted” and distinct voices within institutions in a self-described process of learning and reinvention. Some sharpen their attacks from the outside. And some, too, have looked inward and sought to raise questions—in their practice and in their work—about the once-sublimated political and ethical questions in the arts that are emerging with heightened visibility. French choreographer Jérôme Bel gave up touring by plane in order to lower the environmental impact of his work. His decision sparked a provocative debate about the privileged conditions that make his ethical stand possible (Rodríguez 2021). In 2021, the Yup’ik choreographer Emily Johnson developed and shared online a rider to her presenting contract that requires venues, among other things, to commit to including more Indigenous and BIPOC artists in their programming (see Johnson Reference Johnson2023; ILTA 2023; see also Wingenroth Reference Wingenroth2022).

The principal topics in these examinations have been important and searching. How has the decision about what constitutes aesthetically worthy artistic production been shaped by legacies now regarded as colonial and oppressive? How does the composition and distribution of authority within institutions and artistic fields contribute to aesthetic judgment? Who should art serve, aesthetically or as a social practice? And, who, in turn, should authoritative and mediating artistic institutions serve?

Remarkably absent from this period of ethical and aesthetic self-reflection has been almost any discussion of digital technology and its political implications for artistic production, distribution, and consumption. To varying degrees, both artists and cultural institutions have made use of the technological tools that have now evolved into the creator economy. Most prominently, artists and presenting institutions alike have, over the past decade, been increasing the availability of artistic production available through digital platforms. But neither artists nor presenters have probed the politics of the tools they readily use.

In fact, rather than approaching digital technology with caution, especially as the larger society has begun to cast a more skeptical eye on its potential social harms, the Covid-19 pandemic and associated closures turned a steady stream of digital experimentation into a flood. During the pandemic, venues often migrated live performances and in-person encounters to digital platforms out of necessity, offering live streamed concerts or plays, or virtual art exhibitions. Some institutions have continued to offer digitally mediated productions well past the acute phase of the pandemic. Artists and cultural institutions participating in digitally native experiences, such as building interactivity into performances or “tweeting” about one’s own work, are part of the trend towards increased digital programming. (According to Eichhorn, self-promotion through tweeting builds “content capital,” and is a process of producing content about one’s self to attain and confer cultural legitimacy on one’s creative production [2022:84].)

Typically, the axes of the debate over these forms of creative production are aesthetic (the content is good or bad) and/or managerial (the content is revenue-producing or not). But the question of whether these decisions (“should we put it on YouTube?”), individual or institutional, constitute a speech act for or against a technologically impelled and corporately organized transformation in the creative mode of production is rarely asked.

But we need to ask the question. “Going digital” is not merely a mode shift. Digital creative production is inherently political because it accepts an economic logic that prioritizes the relentless pursuit of attention share over and above any other source of social value associated with aesthetic contribution. To engage in digital creative production is to say something about what art is (content) and is for (attention share).

We believe this introduces at least four vulnerabilities into the fine and performing arts.

Generative Artificial Intelligence

We embarked on this project in the spring of 2022. In a particularly salient illustration of the pace of technological development, it was during the drafting stages of writing this piece that image- generating models like DALL-E, Midjourney, and Stable Diffusion began making waves. Then, on 30 November 2022, the hybrid for-profit/nonprofit organization OpenAI released ChatGPT 3.0, an electrifying development that almost immediately sparked a much wider conversation about the future of so-called artificial intelligence.

During the fall of 2022 and early 2023, we puzzled over whether and how to acknowledge these shifts in the context of this analysis. On the one hand, these tools do not represent the same kind of digital commodity fetishism that we have described. Our discussion has been rooted in the classical Marxian notion that the relations of production are ultimately social, that is, fundamental to the structure of relationships among people in the society. We have been at pains to show how the abstraction of creative expression into a digital commodity impacts the producer and the consumer of the performing arts in concrete, tangible ways. Our concern has been less about the humanistic ineffability of the artistic act than about how the instantaneous and ubiquitous processes of digital commodification threaten to rob audiences access to aesthetically wondrous experiences and to further immiserate artists while also dispossessing them of the full range of creative control. The aesthetic meaning of automated incursions into visual art, speech, and writing, as dispiriting or intriguing as it may be, is not the primary object of our analysis.

On the other hand, generative artificial intelligence systems can hardly be conceived of as discontinuous with the technologies whose development and deployment have animated the transformations we have described.^{Footnote 10} They are, in fact, an evolution of these technologies—if not their apotheosis. Algorithmic systems “smart” enough to create adequately engaging content threaten to render human input into the system we have been describing extraneous.

A few features of these new technologies are particularly pertinent to the issues we have raised.

The first issue is that generative artificial intelligence makes it plausible to imagine a genuinely closed loop in the act of creative expression. Humans remain a pesky inefficiency in almost any system of production, especially systems that trade on knowledge labor. Even the most disempowered worker has basic caloric needs and an efficiency frontier of about half the 24-hour day. One of the productive benefits of digital systems is that they can weather extreme churn. As one content creator burns out, another can take their place. All that matters from the perspective of dynamic growth is a user rate of replacement of x, where x > 1.

Generative artificial intelligence tools have begun to intimate the possibility of the nearly costless production of visual and text-based narratives without the need for human labor at all. As it is, all these tools require at this stage is a human prompt. But it’s not difficult to imagine a scenario in which AI-generated prompts produce AI-generated scripts for AI-generated movies “performed” by AI-generated actors.^{Footnote 11} Arguably, this is the end toward which these systems are directed in the first instance. In an interview on the TWIML AI Podcast, Emad Mostaque, founder of Stability AI, the company behind Stable Diffusion, uses the example of PowerPoint to illustrate how this might work:

In a manufacturing context, the typical defense of “closed-loop” artificial intelligence systems is a conjectural theory of human liberation. As the argument goes, the transition to automation will produce enormous amounts of wealth while displacing huge portions of the human labor force. And since for so many, work is terrible anyway, we can use some portion of that wealth to provide for basic needs, enabling the emergence of a massive leisure class that can direct its productive energies toward self-actualization. This line of reasoning is a big part of Silicon Valley’s seeming obsession with “universal basic income” schemes (Altman Reference Altman2021).

But we should be suspicious of any narrative of human liberation from labor. As noted earlier, the advent of the social web has actually intensified the geographic displacement of human subjection and immiseration that began 50 years ago with globalization. Unlike the sweatshop worker whose body is falling apart making Nike shoes, it is the social media content moderator’s soul that is under threat from wading through the effluvia of human hate and psychopathy. Not to mention, of course, that there are some kinds of work from which we do not seek liberation. Humans do not paint, sing, dance, or write because machines can’t. We do these things because we enjoy them.

Perhaps more worrying, the rise of generative artificial intelligence suggests that the “liberation” of the artist can only be achieved by a stunningly comprehensive act of creative expropriation. Just where do we think these wondrous tools are learning how to reproduce the magic and craft of visual and literary arts? They are ingesting terabytes of publicly available digital indices of novels, poems, and visual arts as well as scooping up proprietary sources like Getty Images (Chayka Reference Chayka2023a). About this, the architects of these systems are glib. Mostaque describes his model as “two billion images, a snapshot of the internet, compressed down” (in twiml 2022:15:10). He jokes, “Artists never make money, right?” (14:27). Midjourney’s founder David Holz shrugged when Forbes asked if his company sought consent from living artists. “We weren’t picky,” he explained, meaning they just took everything they could find (in Salkowitz Reference Salkowitz2022).

Along with many other vexatious questions regarding these new technologies, courts have only just begun to untangle whether what these companies are doing is legally impermissible, either because it constitutes copyright infringement, or violates the DMCA or other statutes. And legislators will have to decide if they want to regulate this area and how. But we should not expect either the courts or Congress to resolve matters in a way that prioritizes the needs of independent artists, or on a schedule fast enough to prophylactically protect them.

A second issue is the homogenization these tools produce. While Holz asserts that “we see this technology as an engine for the imagination,” rather than its demise, elsewhere he acknowledges the model’s humdrum “default styles,” admitting the artificial intelligence program “likes to use teal and orange” (in Vincent Reference Vincent2022; see Herrman Reference Herrman2022). As discussed earlier, “democratization” of the arts merely means a legion of creators contributing within a predetermined set of aesthetic parameters, feeding audiences through hyperpersonalized discovery tools that, paradoxically, guide everything they view toward a bland mean. Are we ready for a world in which aesthetic horizons stretch all the way from teal to orange?

A third and related issue is that all of this happens in a context in which the generative artificial intelligence tools obfuscate the mechanics of creative and aesthetic production, absolving the technology operators of responsibility both for the creative outputs and for the impact of their services on existing economies of aesthetic production. What happens to art when the technology setting the defaults within which “creative” output is produced is described by its own creators as having “no will, it has no goals, it has no intention, no storytelling ability” (Vincent Reference Vincent2022)?

Generative artificial intelligence takes abstraction of creative expression to new heights, or perhaps new depths. Original artistic production is absorbed into a dataset used to train an automated engine of creative production, which is activated through a series of prompts. The prompts trigger the creation of outputs that cannot be audited, explained, or understood in any meaningful sense by the person keying in the prompt or even by the developer of the model. What’s more, having absorbed the “essence” of creativity—here defined as a mathematical representation of the common features of the billions of images in its dataset—the generative AI system divorces the idea of the artist from the actual creation of art. It is the person keying in the prompt who now calls on the algorithm to “produce” art. To wit, there’s already a term, “prompt crafting,” for the skill of manipulating these systems to get a desired result. Companies are recruiting this new “expert,” offering mid–six figure salaries, far more than most practicing artists earn (Anthropic n.d.). Who or what is the actual artist in this scenario? According to Chayka, the artist becomes:

Rex Woodbury is sensitive to some of the philosophical questions in play: “There are some dystopian elements to AI’s creative side, of course; at what point does AI stop augmenting our creativity and start negating it?” he asks in a post (2023). Nonetheless Woodbury refers to these tools as “amplifying human creativity,” noting that “what’s so exciting about generative AI; it can make anyone (even me!) more creative.” Woodbury even asserts that “AI is certainly proving that it’s capable of human-level creativity.” The proof? “In one study, researchers compared human-generated ideas with AI-generated ideas across six chatbots, reporting: ‘We found no qualitative difference between AI and human-generated creativity’ (!)” (2023).

Generative artificial intelligence does not “amplify” (or augment, or extend, or approximate) human creativity. It modifies human productivity. Like all tools, AI affects human labor power. I can use Midjourney to render an image I would not be able to produce with my own hand. But that does not make me more creative. The assembly line undoubtedly augments the productive capacity of the individual worker. It does not amplify their humanity. Quite the opposite.

The designers and operators of these systems seem to be aware of this, though without really understanding the implications. “Another advantage for artists,” says Holz, “is it gives people confidence in areas they’re not confident in. Most if not all artists feel like there’s some part of art they can’t do well” (in Salkowitz Reference Salkowitz2022). In other words, generative artificial intelligence saves the artist from having to work on their craft. Writer Ted Chiang recently made this point about ChatGPT and other large language models:

That dissatisfaction drives the creative process, and brings with it other benefits, in the form of humility, experience, and, eventually one hopes, mastery.

W.H. Auden once enjoined of artists that “Insofar as poetry, or any other of the arts, can be said to have an ulterior purpose, it is, by telling the truth, to disenchant and disintoxicate” ([1962] 1963:27). To the extent that generative AI, even more so than other digital creative tools, secrets away deep within its neural networks the link between the creative impulse and the final output, it actually denigrates creativity, recreating in an ersatz form a pale, homogenized imitation of the human imagination. Creative expression that we cannot explain—or decompose, or understand—is all enchantment, all intoxication.

The answer to Woodbury’s rhetorical question, “at what point does AI stop augmenting our creativity and start negating it?” is: at inception.

What the “Creator Economy” Reveals about Dominant Digital Logics

The creator economy is important not just for its impact on the arts, but also for how it illuminates the impact of internet-enabled technology on larger questions of political economy. As noted at the outset, concern over the role of digital technology in our social, economic, and political lives has reached a crisis point. Our anxieties encompass every conceivable kind of socioeconomic distress, from individual health and well-being to the society-wide collapse of any semblance of epistemic consensus. ^{Footnote 12} What typically links these critiques of digital technology is a recognition that particular features of digital technologies have, variously, exacerbated existing individual and social pathologies, accelerated their spread, or transmuted these pathologies into something qualitatively worse (or all of the above). The picture is one of metastasis: whatever was already there is worse and is spreading. The prognosis is grim, probably terminal.

Since the “techlash” in the wake of the 2016 election, a growing number of theorists, scholars, technologists, and cultural commentators inquired into the features of digital technologies that distinguish them from earlier technology, the forces that produce such pernicious effects, and what can be done to ameliorate our relationship with technology to reap its benefits without enabling its pathologies.

The dominant analysis has been focused on the relationship between these technologies and human attention. ^{Footnote 13} The specifics vary, but the basic argument is as follows:

• Digital services directly and indirectly generate revenues from attention. Direct attention- generated revenue comes from digital advertising, while indirect attention-generated revenue comes from converting attention into a transaction for goods or services.

• Generating attention digitally requires historically low marginal costs. Vast markets can be reached instantaneously and with minimal labor and capital input.

• Such attention is, however, difficult to sustain. As a result, digital services must manipulate human attention in order to (1) acquire a greater share of any one hypothetical customer’s attention and (2) increase the total share of attention available.

• In essence, these services compete in an attention market. They want as much of the time you already spend on your device as they can get, and they are working to keep you on your device longer and longer.

• The imperative of this attention economy is, therefore, to increase the availability and acquisition of human attention at all costs.

This analysis persuasively explains many of the essential mechanics of digital businesses, both those that are dependent on advertising and those that sell goods and services by competing through their “social” features (the ability to “like,” “share,” “follow,” generate new content, and so on). It also appeals to our moral outrage with regard to the negative consequences of digital technology, provoking indignation at the psychological techniques used to manipulate our attention by modifying our behavior.

But one of the consistently overlooked issues in this transformation—one brought into clear focus by the “creator economy”—is the way digital services operators changed the categories of social and economic transactions. In a sense, this is their greatest invention. Facebook redefined “community” as an algorithmically managed digital information and content network and “friends” as users of this network. ^{Footnote 14} Uber redefined “vehicle for hire” services (taxis) as a technology service whose customers are both riders and drivers. ^{Footnote 15} Airbnb performed the same trick with “lodging.” ^{Footnote 16} Each one of these businesses proceeded to devour the markets of leading incumbents, all while vociferously protesting to regulators (and other critics) that they weren’t in the same business at all (i.e., Facebook: “We are not a publisher”; Uber: “We are not a taxi service”; Airbnb: “We are not a hotel operator”).

What is so insidious about these semantic maneuvers is that they divorce services from the public benefits with which they have traditionally been associated and regulated. So when executives and founders in Silicon Valley talk endlessly about removing what they call friction, they are referring not only to the brute operating efficiency of a service, but also to the efficiency—or rather inefficiency—of ethical constraints. When Uber displaces taxis, the dispatch system isn’t the only thing they wipe out. They also evade institutional protections for drivers as labor and passengers as customers. The obsessive talk within technology companies about “governance” with regard to their services is little more than an analytically and bureaucratically convoluted exercise to reinstate the moral and ethical considerations they had previously eradicated. Governance discourse is intended to delude people—or perhaps to convince the companies themselves—into believing that the only legitimate form of restraint can come from within.

The creator economy—and the elision of “artist” and “creator”—is the latest category shift that is fundamentally reshaping not just the economic structure of markets, but their social and political character as well. Web 2.0 social media changed how we consume information. What many of us didn’t realize is that changing the mode of information transmission also meant changing what counts as knowledge.

We can also see how the creator economy is changing what is art. As discussed here, it may just change art “all the way down,” to the point that what constitutes a creative act is shaped from conception by digital business imperatives and operated “autonomously” through generative digital neural networks. Confronting this still nascent rupture in the arts can help us to confront the chasms that we face when it comes to knowledge and social experience—two broad dimensions of human existence that already have been profoundly altered.

Where Do We Go from Here?

We believe the time is overdue for introspection and discussion within the arts and culture community about the digitization of the arts, and within the growing community of socio-technical researchers and practitioners about what the arts can teach us about the consequences of digitization more broadly. Here are some questions to get started:

For artists and arts-presenting organizations:

• What will enable greater visibility of digital capitalism in the arts? How can we encourage and strengthen “class consciousness” in cultural production?

• Can the productive and potentially inclusive possibilities of digital technology be harnessed without further entrenching an extractive digital political economy?

• How can aesthetic discernment be preserved without aiding and abetting historic systems of aesthetic exclusion?

• What are morally appropriate forms of engagement with technology operators? When do operators cross the line from having different but potentially complementary incentives to becoming the Sacklers of digital pollution?

For scholars and critics:

• Have scholars and critics been missing the forest for the trees? That is, have they been too focused on the individual sites of digital discontinuity (information, political organizing, commerce) that they have failed to take note of the larger digital capitalist process, which is agnostic to systems of classification embedded within contemporary society?

• Have scholars and critics accepted the anti-institutionalism of digital capitalism as inevitable or even desirable, thereby ceding to digital capitalism the final—and often triumphal—word on institutional norms?

• Can the traditional institutionally located function of aesthetic discernment exist in the digital political economy?

Perhaps most fundamentally, how can we challenge at inception the changes in social categories promoted by digital platforms? That is, stop these changes before they are inscribed in the economy? In other words, how can we move more slowly and avoid breaking quite so many things?