13.1 Introduction: What Is the Extreme Public Sphere?

Violent events involving the far right, including the Charlottesville Unite the Right rally, the Pittsburgh synagogue attacks, and the Christchurch Mosque shootings, led to the removal of several thousands of accounts from some of the most popular online platforms such as Facebook, Instagram, and YouTube. Prominent far-right personalities such as Alex Jones and his Infowars, Milo Yiannopoulos, and Paul Joseph Watson were removed from Facebook, along with Stephen Yaxley-Lennon (known as Tommy Robinson) and other accounts associated with the British far right.Footnote ¹ These removals peaked in January 2021, after the US Capitol attack, in a process referred to as the ‘great deplatforming’,Footnote ² when major platforms removed thousands of accounts, including those of Donald Trump himself.

This ‘deplatforming’ took place in an attempt to limit disinformation and incitement to violence on mainstream platforms such as Facebook and Twitter (now X). However, research suggests that while deplatforming improved the quality of information and communication on these main platforms, it also led to increasing numbers of users joining non-mainstream platforms and to the creation of what we refer to here as the extreme public sphere: a collection of smaller ‘alternative’ platforms that emerged as an infrastructure for the far right and other extremist movements that find themselves excluded from the mainstream.Footnote ³ This extreme public sphere is comprised of platforms that adopt a radicalised version of absolute freedom of speech and for this reason intervene only minimally, if at all, in the content produced by their users. It includes platforms such as Gab, Bitchute, Rumble, and Gettr, which were created by and for the far right, but also ‘co-opted’ platforms such as Telegram, Dlive, and Discord which have both mainstream and far-right users. Collectively, these platforms are referred to as ‘Alt Tech’. We understand Alt Tech as a political media infrastructure for the far right that supports and sustains the extreme public sphere. Research has found that Alt Tech platforms such as Gab and Bitchute contain toxic content,Footnote ⁴ including hate speech and disinformation.Footnote ⁵ In this context we ask: is the new regulatory approach of the EU sufficient to address and regulate the extreme public sphere?

In focusing on the extreme public sphere this chapter seeks to identify the regulatory challenges that its existence poses specifically for European policy and regulation of digital media. It begins with a discussion of the specificities of the extreme public sphere and an overview of relevant research findings (Section 13.2). Section 13.3 looks into regulatory developments in Europe: in particular the evolution of platform governance, which combines the voluntary Code of Conduct on countering illegal hate speech onlineFootnote ⁶ and the Strengthened Code of Practice on disinformationFootnote ⁷ with the legally binding requirements of the Digital Services Act (DSA).Footnote ⁸ Section 13.4 identifies three interrelated challenges that the extreme public sphere poses for these regulatory instruments: (i) they are not motivated to act in ways that the regulatory instruments assume, as they are political and not only economic actors; (ii) their diverse business models do not rely on advertising; and (iii) their minimal obligation under the DSA is to remove illegal content when it gets reported to them.Footnote ⁹ One of the aims of the DSA is to protect and safeguard the operations of a pluralistic but democratic, inclusive, and respectful digital public sphere, balancing freedom of expression and the right to non-discrimination, as alluded to in Recital 3; the operation of Alt Tech presents a serious challenge to achieving this aim. Section 13.5 of this chapter brings all the threads together and discusses a potentially more fruitful avenue to follow for addressing the extreme public sphere.

13.2 Understanding Alt Tech

Research on Alt Tech and the broader information ecosystem has provided crucial insights into the rise and operation of this new breed of platform. The general agreement is that although smaller and fringe platforms have existed for quite some time, Alt Tech as it appears today emerged in the context of the mass deplatforming of far-right and other extremist accounts from the big three platforms: Facebook, YouTube, and Twitter. This deplatforming led prominent accounts and their followers to migrate to smaller platforms or build their own: Alt Tech is therefore seen as an alternative media infrastructure supporting extreme ideologies.Footnote ¹⁰ The key characteristic of Alt Tech is a lax approach to content moderation, with an emphasis on an understanding of freedom of expression as almost completely unrestricted speech. Given this absolutist freedom-of-speech approach, it is not surprising that the content they host contains a large volume of hate speech, dis- and misinformation, and other kinds of toxic and problematic speech. Finally, an important attribute shared by most Alt Tech platforms is that they do not have data-intensive business models that rely on the collection of data from users. This in turn is directly connected to their absence of elaborate algorithms and recommendation systems for structuring and guiding the user experience. As we will argue later, this is an important difference between Big Tech and Alt Tech and constitutes one of the challenges for policy.

A series of high-profile violent events driven by the online far right led to increased pressures on platforms to act. Events such as the Unite the Right Rally, the Pittsburgh synagogue terrorist attack, the Christchurch mosque terrorist attack and the events on 6 January 2021 in the US Capitol all contributed to the mass deplatforming of the far right, including several highly prominent accounts, such as the account of former US President Donald Trump. In his groundbreaking article, Rogers studied the effects of deplatforming by tracking the activities of ‘extreme celebrities’, finding that Alt Tech constitutes a viable destination for them even if they have a smaller fanbase there.Footnote ¹¹ Rogers further reports that several of these celebrities of the far right direct users to their personal websites rather than to other social media platforms.Footnote ¹² In their study on deplatforming, Ali and colleagues report similar findings: deplatforming results in a migration to alternative platforms, diminishing the reach of those deplatformed but often accompanied by a heightened volume of toxicity.Footnote ¹³ In sum, there is general agreement that deplatforming (i) leads abusive users to Alt Tech platforms, which (ii) diminishes their reach but (iii) may radicalise those users further, pushing them to more extreme content; additionally, (iv) there is some evidence that extreme users rely on Alt Tech for community-building and use mainstream platforms for propaganda and recruitment.Footnote ¹⁴

But deplatforming further led extreme actors to the realisation that dependence on mainstream digital platforms entails the possibility of being entirely cut off from their actual and potential audiences. This threat to their survival, argue Donovan, Lewis, and Friedberg, pushes the extreme right towards creating their own infrastructures.Footnote ¹⁵ Donovan and colleagues view the rise of Alt Tech as a form of tactical innovation in response to the calls for ‘no platforming’ for far-right actors, but also enabled some of them to coalesce, organise, and amplify their voices.Footnote ¹⁶

For Alt Tech to succeed as tactical innovation for the far right, it had to not only fulfil its purpose of providing a means for identity-building, organisation, and propagation of the ideas of those movements; it also had to ideologically align with the far right. This is reflected in a dual move: appropriating the norm of freedom of speech and giving it a very specific meaning, that of the removal of any limits to the circulation of content; and putting this at the centre of both their overall ideological approach and their new Alt Tech infrastructure. Donovan and colleagues quote Gab’s announcement of the so called Alt-Tech Alliance:Footnote ¹⁷ ‘The Free Speech Tech Alliance is a passionate group of brave engineers, product managers, investors and others who are tired of the status quo in the technology industry. We are the defenders of free speech, individual liberty, and truth’.Footnote ¹⁸ While in practice Alt Tech platforms have to impose some limits, if only so that their apps can be listed in app stores, this version of absolute freedom of speech animates all of them.

Alt Tech’s combination of absolute freedom of speech as its key value, non-existent or rudimentary moderation practices, and the extremist ideology of far-right groups, which includes white supremacy and deep-seated misogyny,Footnote ¹⁹ means that Alt Tech is characterised by high volumes of hate speech, disinformation, conspiracy theories, and other kinds of toxic content. Comparatively speaking, Zannettou and colleagues report that Gab has a much higher prevalence of hate speech than Twitter/X but contains less hate speech than 4chan/pol/.Footnote ²⁰ Using a similar methodology, Trujillo and colleagues found that Bitchute contains more hate speech than Gab, but still less than 4chan.Footnote ²¹ There is evidence to suggest not only that these kinds of discourses are widespread on Alt Tech platforms but also that they can become increasingly extreme. Dehghan and Nagappa found that content on Gab about vaccination was already verging on the conspiratorial, but as time went on it intensified and became more explicitly political.Footnote ²² For Dehghan and Nagappa this question of depth or degree of extremism is an important element of Alt Tech that should be studied in addition to the spread of extremist content and narratives.Footnote ²³

A final important attribute of Alt Tech platforms concerns their business model and associated internal organisation. Most Alt Tech platforms do not rely on the data-intensive model that is central to the very large mainstream platforms. Few, if any, collect any user data at all. Some Alt Tech platforms, for example Gab and Bitchute,Footnote ²⁴ adopt a clear ideological stance in favour of free information flows and decentralisation and, therefore, they do not enclose users in ‘walled gardens’ to collect their data. Others, for example Telegram, put user privacy at the centre of their model. Generally, most of these platforms have two sources of revenue: premium services, for which they charge a monthly subscription (e.g. Telegram and Rumble business accounts), and/or content monetisation, whereby they enable and support content creators in monetising their content so that they can then receive a fee for hosting them (e.g. Bitchute).Footnote ²⁵ Because Alt Tech platforms do not generally collect user data, their algorithms for ordering content tend to be much simpler than those of very large online platforms. On Bitchute and Rumble, which are both video platforms, users viewing content made by a creator will not get recommendations for similar creators or content, as on YouTube, but will get more of the content by the same creator. Any recommendations will be on their landing page and will depend on a mixture of recency and number of views.

To summarise, Alt Tech platforms differ substantially from mainstream digital platforms in their ideological affiliation with the far right, in their adoption of absolute freedom of speech as their key value, and in their business models, which do not rely on user-data extraction or advertising revenues. In this sense, Alt Tech can be seen as a political media infrastructure that sustains the far right and related ideological groupings.Footnote ²⁶ This political function of Alt Tech presents significant challenges for current platform regulation models.

13.3 Platform Regulation

Approaches to platform regulation can be broadly classified in terms of four ideal-typical approaches: (i) authoritarian; (ii) libertarian; (iii) liberal; and (iv) critical.Footnote ²⁷ While these ideal types reflect different underlying values and ideologies, existing regulation combines them, borrowing some elements and discarding others. This section will first present these ideal types and then discuss the European Union’s approach to platform regulation by discussing its Codes of Conduct and Practice and Digital Services Act. It will argue that the EU’s approach borrows from the critical approach in its concerns over data extraction and surveillance capitalismFootnote ²⁸ but for the most part follows a moderate liberal path of co-regulation combined with elements of libertarianism, viewing platforms as rational economic actors that have a stake in regulating themselves.

In her genealogical approach to platform governance and more broadly to internet regulation, Bietti identifies three broad ideological positions: anarcho-libertarianism, liberalism/neoliberalism, and critical conceptions.Footnote ²⁹ Libertarianism is associated with early internet techno-optimism and is typified in John Perry Barlow’s cyberlibertarian manifesto, ‘A Declaration of the Independence of Cyberspace’.Footnote ³⁰ Under this perspective, regulation stifles innovation and state interference is entirely unwelcome. For proponents of this view, the internet is capable of self-governance and it should be left to its own devices. This position constituted more of a utopian techno-optimist narrative, characteristic of the early days of the internet, rather than a fully-fledged view on regulation.

While the internet developed initially without any regulatory interventions, liberal critiques pointed out that the lack of regulation does not guarantee individual freedom. For critics such as Lawrence Lessig, computer code was already structuring cyberspace and in this sense was acting as law.Footnote ³¹ In the absence of regulation, users/citizens are unprotected and their ability to use the internet can be compromised. From the liberal perspective, regulation is seen as imposing some limits in order to protect individual liberty. Bietti points out that such liberal views tend to oversimplify the issues involved, considering regulation a matter of planning and development of specific technical protocols.Footnote ³² In focusing on these, liberal views eschew questions of power distribution and unpredictable risks. This is where the critical perspectives come in. For such perspectives, technologies are not neutral artefacts but embed power relations and can also shape political realities.Footnote ³³ Critics such as Yochai Benkler proposed a peer-to-peer, commons-based governance that would address questions of emerging power dynamics and enclosures on the internet.Footnote ³⁴ Others, such as Barbrook and Cameron, made an explicit connection between the hippy libertarian ideology and the neoliberal approach to markets and the economy. They posited the emergence of a new ideology, the Californian Ideology, specific to the internet and characterised by techno-market solutionism, limited state involvement, and freedom understood in terms of individual preferences.Footnote ³⁵

Bietti convincingly argues that these ideological frameworks underpin current platform regulation.Footnote ³⁶ Similarly, Viejo Otero has shown that there are three approaches to content moderation on digital platforms. The first, freedom-of-speech absolutism, is more or less what we have already identified with respect to Alt Tech.Footnote ³⁷ Second, the liberal position, which is currently in place on large platforms such as Facebook, Instagram, and YouTube, assumes a position of neutrality; some critics have found this to be essentially blind to the realities of racism.Footnote ³⁸ Third, the social justice approach found on some Mastodon servers approaches content moderation from the point of view of silenced communities, not only protecting them but also in certain contexts prioritising their voices.Footnote ³⁹ To these we may add an authoritarian position, which we can associate with the policy of imposing tight controls on social media content that is found in some states, notably Iran and China.Footnote ⁴⁰

It is this context that saw the emergence of the DSA, the EU Code of Conduct on Countering Illegal Hate Speech Online, and the EU (Strengthened) Code of Practice on Disinformation. These measures represent a distinct European approach that combines various elements from these ideological positions. For Viejo Otero, the European approach is a stronger version of the liberal view:Footnote ⁴¹ it seeks to reconcile freedom of expression, enshrined in Article 11 of the EU Charter of Fundamental Rights,Footnote ⁴² with protection of minorities, recognising specific historical events such as the Holocaust. Like the liberal position, the European approach prioritises the rights of individual users, and in particular their right to freedom of expression and information as found in Article 11 of the Charter and the freedom to conduct a business as found in Article 16. But crucially, the European approach views co-regulation as the optimal perspective,Footnote ⁴³ displaying affinities with both the libertarian and the (neo)liberal positions.

In particular, while the DSA is legally binding for EU Member States, the Codes include both self- and co-regulatory approaches. In self-regulation, such as in the Code of Conduct on Countering Illegal Hate Speech Online, digital platforms that have opted in volunteer to monitor and enforce the Code. In co-regulation, such as in the Strengthened Code of Practice on Disinformation, the European Commission monitors the enforcement of the Code on a regular basis. From the point of view of the European Commission, the Codes constitute a set of expectations that signatories agree to meet. These expectations are important in setting the broad parameters of how to tackle illegal and/or problematic content such as hate speech and disinformation. The Strengthened Code of Practice on Disinformation, in particular, offers a comprehensive set of policies covering (i) demonetisation of disinformation; (ii) labelling of political adverts; (iii) preserving the integrity of the services by removing bots, fake accounts, and the like; (iv) empowering researchers by offering access to data and users through labelling and literacy; (v) monitoring and transparency measures.Footnote ⁴⁴ The Codes clearly recognise that social media corporations can and should do a lot more to keep their platforms free of illegal hate speech and disinformation; while the Codes are nominally voluntary, they are increasingly concretised as formal obligations and constitute a form of soft law.Footnote ⁴⁵

In the binding regulatory part, the DSA steps up these efforts. It begins by creating a tiered system, distinguishing between services considered to be ‘mere conduits’, ‘hosting’, and ‘caching’,Footnote ⁴⁶ and online platforms, which not only host but also disseminate information.Footnote ⁴⁷ Online platforms in turn are divided into two categories. The first comprises very large online platforms (VLOPs) and search engines (VLOSEs), designated as platforms that have more than 45 million users per month in the EU; the other category comprises platforms with less than 45 million users. While all platforms are subject to the same baseline obligations, VLOPs and VLOSEs have additional, more stringent obligations. These additional obligations are found in Section 5 of the DSA.Footnote ⁴⁸ For the purposes of this chapter we focus on three categories of obligation, which illustrate most clearly the challenges raised by Alt Tech platforms: risk management; transparency; and independent audits.

For online platforms that are not designated as VLOPs/VLOSEs, there are no specific obligations around risk management and mitigation; this means that by virtue of not being VLOPs/VLOSEs, Alt Tech platforms do not need to assess the risk that their platforms may be hosting illegal content. Recital 104 of the DSA states that ‘risk mitigation measures concerning specific types of illegal content should be explored via self- and co-regulatory agreements’ but does not contain any binding rules.Footnote ⁴⁹ In contrast, VLOPs/VLOSEs have obligations around the management of systemic risks, such as large-scale manipulative behaviour enabled by algorithmic ordering and recommendation systems. They must conduct periodic risk assessments covering risks around the dissemination of illegal content; negative impact on fundamental rights, civic discourse, and electoral processes; and gender-based violence.Footnote ⁵⁰ They must then take measures to mitigate these risks, for example through adapting their terms of service and their algorithms.Footnote ⁵¹ Smaller platforms are exempted from these obligations. Secondly, while all platforms must publish annual transparency reports on content moderation activities,Footnote ⁵² VLOPs/VLOSEs must provide more detailed reports every six months, including information on their risk assessment and mitigation measures.Footnote ⁵³ Thirdly, VLOPs/VLOSEs are required to undergo annual independent audits to assess their compliance;Footnote ⁵⁴ there is no such requirement for smaller platforms.

Further, the DSA retains the same liability regime as the E-Commerce Directive of 2000,Footnote ⁵⁵ where service providers are not deemed liable for illegal content created by their service users when they do not have actual knowledge of such content or when, upon obtaining such knowledge, they act promptly to remove or to disable access to the illegal content. Article 7 stipulates that these intermediary service providers will remain exempt from liability even if they take measures to detect, remove, or disable access to illegal content, while Article 8 makes clear that these platforms are not under any general obligation to monitor the information they transmit or store. Their obligations begin when they receive orders from relevant authorities to take action against illegal contentFootnote ⁵⁶ and to provide information about individual users, where this has a legal basis.Footnote ⁵⁷ Compliance with the DSA is mandatory, and breach of the rules may lead to hefty fines of up to 6 per cent of the annual worldwide turnover of VLOPS/VLOSEs.Footnote ⁵⁸

There is little doubt that the DSA constitutes a groundbreaking development in the field of digital policy and regulation. Combining self- and top-down regulation and adopting a tiered approach, separating the obligations of VLOPs/VLOSEs from those of smaller platforms, it aims to occupy the medium ground between liberalism and libertarianism, safeguarding freedom of expression while protecting citizens from the pernicious effects of hate speech and disinformation. Moreover, it still affords platforms a significant say in their own regulation. This is because, despite having more obligations under the DSA, platforms still retain a significant degree of autonomy in developing and implementing internal policies for the flow of content on their outlets, selecting trusted flaggers, performing monitoring exercises, and structuring their transparency reports. Finally, platforms still enjoy a degree of protection from liability, since the DSA upholds the liability regime of the E-Commerce Directive.

Importantly, the DSA is essentially a continuation of the E-Commerce Directive and operates as a form of market regulation, looking to harmonise rules concerning platforms while also protecting users/consumers from illegal content.Footnote ⁵⁹ In this context, platforms are considered rational economic actors that are motivated to act in ways that protect the integrity of their services and thereby protect their users from harm. Platforms are motivated to adhere to regulation because it helps them create a trustworthy and ‘clean’ environment for their service users, including advertisers, small businesses, and individual users. Since both platforms and regulators want to ‘clean up’ platforms, phenomena such as hate speech and disinformation are seen as containable through technological means. These include addressing potential side-effects of their internal design, algorithms, and affordances, and other types of interventions such as, for example, the use of AI to remove or demote certain content and the reliance on trusted flaggers and digital literacy initiatives. While this is likely to hold true of the VLOPs, it is not the case with Alt Tech platforms. Section 13.4 addresses the challenges that Alt Tech poses for the DSA.

13.4 The Challenges of Alt Tech

As we have seen, Alt Tech consolidated its presence as a political media infrastructure in the aftermath of far-right activism and processes of deplatforming that purged far-right extremist accounts from the main digital platforms.Footnote ⁶⁰ Alt Tech emerged as a tactical innovation to enable far-right movements to continue their activism and propagation of their ideas.Footnote ⁶¹ Its internal structure and design features differ substantially from those of the main digital platforms and are geared towards supporting content creators, who tend to be far-right influencers, with minimal intervention in terms of moderation.Footnote ⁶² Since Alt Tech platforms are in the first instance oriented towards a very specific public, their revenue model is substantially different from that of the main platforms, as they do not collect data for microtargeting. The challenges that Alt Tech poses for the DSA and its approach to regulation stem from these characteristics. Based on these, we can identify the following challenges. Firstly, Alt Tech platforms are not only or primarily economic but also political actors and have altogether very different motivations from those of mainstream digital platforms.Footnote ⁶³ Secondly, as of 16 August 2024, no Alt Tech platform has been designated as a VLOP or VLOSE and therefore none of them has any obligation to audit algorithms or identify systemic risks.Footnote ⁶⁴ Thirdly, their business model – supporting content creators through micropayments but also venture capital, donations, and subscriptions – makes them unaccountable to advertisers.Footnote ⁶⁵ Finally, Alt Tech platforms such as Bitchute nominally comply with requirements to have a content moderation system in place and to publish transparency reports.Footnote ⁶⁶ But this nominal compliance does not mean that the platform is free of hate speech and disinformation. This section elaborates on these challenges and their implications.

The brief exposition of the DSA and the Codes in Section 13.3 of this chapter made the point that these regulatory instruments consider digital platforms as rational economic actors, motivated by profit and having a stake in a ‘clean’ platform in order to attract and retain advertisers. Indeed, the advertiser exodus observed on Twitter/X under Elon Musk shows that most advertisers are looking for properly moderated platforms.Footnote ⁶⁷ But Alt Tech platforms are not in the first instance acting as economic actors appealing to advertisers. Rather, they operate as political actors motivated by ideology and are addressing a specific category of users who share their worldviews. A central value across all Alt Tech platforms is that of a particular version of free speech, understood as almost unhindered information flows. For example, Gab describes itself as ‘a free speech software company’Footnote ⁶⁸ and Rumble sees itself as ‘immune to cancel culture’.Footnote ⁶⁹ This commitment to free speech is not entirely a political one: it is also a means by which they differentiate themselves and attract and keep a specific community of users. As Gillespie has argued, moderation is the commodity that platforms sell,Footnote ⁷⁰ and this also applies to Alt Tech; in their case, however, they are selling non-interference, minimal or non-existent moderation of content. Since their brand identity is directly entangled with minimal moderation, it is very difficult to see how they could adhere to the spirit if not the letter of the DSA.

The DSA and the Codes expect and stipulate that platforms have a content moderation system in place in order to remove illegal content or content that goes against their terms of service. While some VLOPs have automated systems, smaller platforms rely almost exclusively on user reports. Some Alt Tech platforms, most notably Bitchute and Telegram, have content moderation systems in place where users can flag content for, among other reasons, terrorism, dogpiling, spam, fake accounts, violence, pornography, child abuse, and copyright violations. Bitchute geo-blocks content that is illegal in Europe: for example, incitement to racial hatred or Holocaust denial. But both platforms rely on user reports, and their users tend to agree with the viewpoints and politics of most of the accounts/channels on such platforms and therefore are unlikely to report any but the most problematic content. Indeed, in its 2023 Transparency Report, Bitchute reported that in the first quarter of 2023 it received 2,474 reports for incitement to hatred (which is illegal in the UK, EU, and EEA territories); of these it found that 1,650 were valid reports and actioned them, affecting 11,231 videos.Footnote ⁷¹ Bitchute calculates the prevalence of this kind of content at 0.0847 per cent. This does not account for the findings of Trujillo and colleagues that Bitchute contained 4.4 times the number of hate terms encountered on Twitter/X.Footnote ⁷² While Alt Tech platforms have a content moderation process and meet to a degree the formal requirements of the DSA, they are still likely to host a large volume of problematic content. It can be said that they meet the letter of the DSA in that they have content moderation systems in place, but they do not meet the spirit of the DSA in creating a safe space that upholds fundamental rights.

Even if Alt Tech platforms have some content moderation processes in place, they are under no obligation to audit or check their systems for systemic risks as they are not considered VLOPs. Systemic risk audits for VLOPs are meant to reveal whether their algorithmic systems end up unintentionally diffusing disinformation. In general, platforms’ algorithmic organisation revolves around their profit-making function and is structured in ways that maximise benefit for advertisers while preserving the integrity of the platform’s services.Footnote ⁷³ Because Alt Tech does not rely on advertisers, its algorithmic organisation tends to be different from that of very large online platforms. As already explained, YouTube tends to recommend ‘similar videos’ based on those a user has interacted with before. Bitchute, by contrast, recommends only more videos from the same creator, which then supports creators in building their profiles.Footnote ⁷⁴ Since Bitchute creators deal primarily in far-right content, this algorithmic system allows them to build a following. While VLOPs are both motivated by their economic activities and regulated by the DSA to control for the popularisation and diffusion of harmful content, Alt Tech platforms rely on providing a service to the ‘ideology entrepreneurs’ of the far right and other extremist movements.Footnote ⁷⁵ The architecture of Alt Tech platforms and their internal organisation through algorithms reflect precisely this service that they provide to influencers of the far right. While VLOPs and VLOSEs are subject to annual audit obligations,Footnote ⁷⁶ no audit can lead to any corrections in the case of Alt Tech since their system is calibrated to meet the needs of this particular ideological milieu.

As also mentioned, the business model of most Alt Tech platforms reflects this central service as a political media infrastructure for the far right: rather than relying on advertising income they rely on subscriptions, donations, and micropayments and in some instances on venture capital. For example, Rumble has received funding from Peter Thiel, the chairman of Palantir and a major donor to Donald Trump’s 2016 campaign.Footnote ⁷⁷ These platforms are therefore accountable not to mainstream advertisers but to these ideological groupings of users, channels/creators, and financial backers.

Finally, the DSA obliges platforms to remove illegal content and the Code expects platforms to improve access to authoritative sources and demote disinformation,Footnote ⁷⁸ which Code signatories are likely to do because they have a stake in providing a safe and trustworthy platform to their users and advertisers. However, as we have argued, Alt Tech is providing platform services to far-right users and publics who are producing and disseminating exclusionary and discriminatory narratives against certain groups. Even if Alt Tech platforms remove illegal content when it is reported to them and geo-block overtly discriminatory content, as Bitchute does, their users and channels are quick to adapt their narratives so that they are framed as debates, using metonymies, memes, symbols, and other ways to ‘hide in plain sight’, as May and Feldman put it.Footnote ⁷⁹ The combination of Alt Tech’s operation as a political media infrastructure for the far right and its users’ political views make it very difficult if not impossible to see how the DSA and its co-regulatory approach might work.

The implications of all this include the bifurcation of the public sphere, with the mainstream covered by the DSA and the Codes and an extreme public sphere that can remain almost untouched by regulation provided it meets the minimum requirements. As we have argued, this is because in the DSA, platforms are obliged to take down illegal content only when it is reported by users. Although most VLOPs/VLOSEs are motivated to keep their platforms ‘clean’ and safe to ensure advertiser income, Alt Tech platforms do not share the same incentive; and their users are unlikely to report content they agree with, for example incitement to racial hatred, even if they are illegal. Since these two parts of the digital public sphere are still connected like two communicating vessels, the extreme public sphere will continue to feed its content into the mainstream.

13.5 Conclusion: Is Platform Regulation Enough?

This chapter has focused on the operation of platforms known collectively as Alt Tech. We understand Alt Tech as a political media infrastructure for the far right that supports and sustains the extreme public sphere. The question that animated this chapter concerns the extent to which the DSA is in a position to address the challenge of these platforms. Our analysis of the DSA highlighted the co-regulatory approach taken by the EU and the central understanding of platforms as rational economic actors, motivated to keep their platforms as free as possible from toxic content. This chapter’s discussion of the challenges of Alt Tech has sought to demonstrate that Alt Tech platforms act mainly as political actors, catering to the needs and values of users who are ideologically aligned with the far right and conspiratorial worldviews and who prioritise the value of absolute freedom of expression. Their business model makes them accountable not to advertisers but to these political publics, and they are therefore not sufficiently motivated to design and implement a fully functional content moderation system.

From this analysis we concluded that it is unlikely that the DSA and the Codes can regulate the extreme public sphere effectively. Moreover, it is unlikely that the EU can produce a more stringent form of regulation without contradicting its commitment to fundamental rights and values, such as freedom of expression. The co-regulatory approach is consistent with the liberal ideological commitments and values of the EU, but it inevitably runs into trouble when it comes to regulating the extreme public sphere.

We are not suggesting here that stricter regulation will resolve the challenges posed by Alt Tech. The political character of Alt Tech, and the increasing mainstreaming of some of its far-right narratives,Footnote ⁸⁰ reveal that these challenges are unlikely to be resolved only at the level of regulation. Rather, a whole-society approach may be necessary to tackle the proliferation of hate speech, conspiracies, disinformation, and all kinds of extremist content. By a whole-society approach we mean an approach that addresses racism, misogyny, and other forms of structural discrimination; an approach that recognises and resolves points of social tension and inequalities that feed far-right narratives; and an approach that aims to increase social trust and political efficacy.

Regulatory approaches to platform governance cannot resolve these tensions on their own but should be seen as part of a whole-society approach and should be nested in that. There are important questions, however, regarding the ultimate effectiveness of regulatory instruments such as the DSA. Its stated goals include the provision of a new set of harmonised rules for addressing and removing illegal content while protecting fundamental rights. There are clear references to the protection of citizens’ right to freely express themselves but not many references to the overall health of the digital public sphere and the potential contribution of toxic (but not illegal) content to political polarisation and fragmentation.Footnote ⁸¹ The reluctance of the DSA to name these and its focus on the tip of the iceberg of toxicity (i.e. illegal content) compromises its ability to deliver a healthy digital public sphere.