While the idea of platform governance is not new, the complex nature of its functioning and the reasons for its emergence and entrenchment still lack holistic conceptualisation. While it is impossible to develop such a perspective in a single blog post, the considerations below are intended as a sensitising tool and a call to think about platform governance as simultaneously premised upon societal developments from which it has emerged and a pervasive force shaping contemporary societies.
In order to better understand the how and why of platform governance, at least two arguments are possible, although they are by no means mutually exclusive: a historico-political and an economic one.
On the historico-political side – and for perhaps the most eloquent account, see De Gregorio – attention is drawn to the convergence in time between neoliberal deregulation policies adopted by states and the rapid expansion of the Internet and the associated digital technologies. The companies that have sprung from this development have not only benefitted from the lax regulation of their own business activities but have also stepped in to fill the regulatory and governance lacunae left by the retreating state.
However, there is an important caveat: platform companies have acquired quasi-public functions without the corresponding checks and mechanisms for public oversight. While the public is capable to exert democratic control over decision-making bodies (or, at least, that is the ideal anyway, subsequently enabling critiques of democratic deficit or opening up the space for more radical democratic alternatives) and the regulatory power of states is (or, at least, ought to be) limited by constitutional norms and international human rights commitments, that is entirely absent from platform governance. Crucially, even the normative ‘ought’ element is absent – although platform governance practices are typically premised upon sets of rules, norms, or guidelines, often framed in terms of ‘community’ or similarly popular (if not populist) references, they are, instead, sets of externally imposed demands, based upon the commercial interests of platform companies and/or the personal visions of tech entrepreneurs.
The latter point in particular deserves special attention as the directness, immediacy, and unavoidability of algorithm-based governance (think about the difference between the ability to circumvent the law versus the ability to circumvent coded platform architecture) means that once changes to such commercial considerations or to platform ownership and, therefore, the entrepreneurial vision take place, governance practices can be overhauled with a brutal top-down efficiency that could often make even the most totalitarian of dictatorships envious (think of the speed and one-sidedness of Musk’s overhaul of Twitter). Hence, a clear paradox is evident: platforms and their internal rules acquire quasi-constitutional powers, often balancing fundamental rights against one another, while also redrawing the meaning and scope of such rights, but without the necessary checks and balances. Again, there could be a historical reason: over the last couple of centuries at least, states, and not private companies, were the main actors capable violating human rights and exerting arbitrary power (outside the colonial context, that is). Hence, revolutionary and emancipatory struggles have been primarily directed against the state as a vehicle of oppression. Even when private actors were to blame, this would primarily be due to state capture and not their independent power and capacity. Hence, the safeguards and positive obligations put in place at both national and international level have primarily had the state as their referent, which today is no longer sufficient.
One must also delve more deeply into the value considerations that underpin the mechanisms of platform governance. As stressed above, their normative considerations are very different from those that serve as the normative basis (and legitimation strategies) of regulation by public bodies. Indeed, whereas democratic politics is about articulation of public interests and demands, algorithmic governance is determined by the profit motive and the ensuing business models. Hence, a ‘good’ algorithm makes users engage and, therefore, return to the platform as frequently as possible. As well known, the latter might be contrary to the public good and might even undermine democratic politics – think of, for example, platforms’ willingness to tolerate, or sometimes even encourage, polarisation, conspiracy theories, and otherwise harmful content as long as they generate engagement and, therefore revenue through advertising and data capture or subscriptions (in fact, sometimes platforms would go to great lengths to attract such content, as in Spotify and Joe Rogan). It is here that the second group of considerations (those focused on the economic side of affairs) comes to the fore: as stressed in numerous studies of platform capitalism and as further illustrated by the hype surrounding the latest technology buzzwords, from the metaverse to large language models and other forms of generative AI, the essence of the platform business model is the capture and exploitation of data and of the users as both a resource and a free labour force that becomes indentured through network effects. If the platform can be considered today’s dominant forms of economic organisation, then it is largely due to their capacity to render algorithms into governance devices that structure audience behaviours and data (sur)rendering practices in line with corporate interests.
The power of platform governance rests not only the specificities of code as ‘the law’s new language’ but also on the former operating on the fundamental registers of existence and memory. In terms of existence, platforms become the arbiters of veracity and factuality, often on the basis of AI tools (the training of which remains non-transparent) and human content moderation practices that still rely on the same rules and governance practices that have been shown above to contravene traditional notions of regulation as based on the public interest, however the latter would be formulated and defined. Truth and fact (and their role in public discourse) then become deeply entangled with the interests of platforms and the tech entrepreneurs behind them. An even further step down the ladder of existence as a means of governance would be the threat of de-platforming, i.e. termination of one’s digital existence on the platform: while often framed in terms of noble causes (and occasionally serving them), such as obstructing the spread of hate, terrorist recruitment, child grooming etc., this disciplinary threat further entrenches platform monopoly over truth and legitimate discourse more broadly conceived, while also often disproportionately targeting users who are already underprivileged or whose modes of (self-)expression go beyond the mainstream. Closely connected to that is platform power over memory. The latter includes, among other things, matters of historical record and how they should be remembered and interpreted, such as deciding between historical or cultural record on the one hand and pornography or gore on the other. However, it also concerns the remembering of events and individuals through the control of personal timelines, personalisation of news and other content, and, once again, (dis)allowance of individuals, groups, and organisations on the platform. As such, then, power is exerted trough both reflexive control and imposition of (self-)censorship.
To conclude, societies today are facing a radically transformed reality of platform-centric governance to which traditional tools of defending individuals against external power are ill-suited. While that might, at first sight, appear to be a deeply nostalgic position, pitting an idealised account of the sovereign state against the lawlessness of platform capitalism and a call to ‘take back control’, this is by no means the sole or necessary implication. What should be emphasised is that, instead of reverting to times gone by (after all, the state has rarely been the completely accountable benevolent lawgiver either), we ought to seek new forms of participation and the inclusion of as diverse sets of stakeholders as possible in order to develop a truly inclusive and accountable form of digital constitutionalism.