Regulating the Discursive Power of Big Tech Companies
Big Tech companies have power. One element of this power is discursive power, including in the public sphere, a cornerstone of democratic societies: the place where citizens come together, share information and where public opinion is formed. In the current digitalized society, the public sphere has a significant online component. Discursive power may continue to grow, fuelled by AI developments, unless checked. To shape a possible legal response – we focus on European competition law – requires understanding the complexity of this power. Though competition law is focused on market power, we argue that it can and should have a role to play in curbing discursive power too.
Power and Discursive Power of Big Tech Companies
Consider, first, the power of Big Tech companies. That this power is multifaceted is the central hypothesis of the ‘Modern Bigness’ theory, which teases out the following elements (Gerbrandy & Phoa (2023)): Big Tech companies’ power is not merely grounded in well-known market power, but also in the (equally well-known) economics of network-effects, in data-gathering and having (access to) computational and thus data-processing capabilities, and in being able to sweep up innovation if it occurs outside the existing corporate structures. This power affects the economic (market) domain, but also the political, social and personal domain; and, apart from its instrumental and its structural dimensions, it also encompasses discursive power. We use ‘Modern Bigness’ as a descriptor and suggest it is different from corporate power before digitalization: the scope of big technology companies’ platform ecosystems allows a continuous reach of users with a level of precision that has not been possible before. Moreover, from the perspective of big technology companies, the roles that users hold in the different domains are irrelevant: there is no incentive to distinguish between citizens and consumers.
Second, let us zoom in on the idea of discursive power, the power to exert influence on the way ideas and concepts, and even knowledge and truth, are framed and interpreted. The term ‘discourse’ is used here to indicate the study of language in use in social interactions. Discourse analysis examines the way discursive practices construct a social order and often focuses on power. For example, Foucault uses the notion of “power/knowledge“. Power and knowledge are intimately connected and mutually reinforcing; knowledge is not neutral or objective but shaped by these power relations. Power determines what knowledge is produced, how it is disseminated. Fuchs, in more recent times, focuses specifically on corporate power in relation to politics. It is her taxonomy, that we also use for our Modern Bigness concept, that distinguishes the dimensions of instrumental, structural and discursive power. Discursive power relations do not only manifest as having a dominant substantive voice in terms of the content of a discourse (what is talked about), but also in terms of access to, and form of participant structures (who can talk, and in what way or form).
Thirdly, Big Tech corporations and discursive power. We define discursive power of Modern Bigness companies as that facet of Big Tech’s power affecting and shaping societal and political discourses, ideas and norms. To make this more concrete, it helps to point out some examples. In terms of Big Tech’s discursive power in relation to content, Big Tech’s eye-watering aggregate lobby expenditure of 113 million euros in 2023 indicates its power to shape political (and academic) agendas and assert significant influence over the content of regulations, such as the AI Act. Furthermore, highly advanced algorithmic recommender systems directly impact the way in which Internet users find and access information online, affecting not just how search engines work but also which content is prioritized in search results or on social media feeds. Big Tech’s online content moderation practices further shape online discourse by influencing what is allowed, how it is presented, and what the behavioural norms are within digital communities. In terms of access to the discourse, content moderation policies have led to several widely commented personal bans, such as Trump’s ban from Twitter, but also to concerns about online censorship and ‘shadow-banning’, the suppression or muting of an online user’s content without informing them. In terms of form, research has shown that the ‘affordances’ (technical features that enable or constrain potential behaviour or use) of online platforms such as X (formerly known as Twitter), such as the potential anonymity and the constraints of first 140, and now 280 characters, impact the quality of public discourse.
How Modern Bigness’ Discursive Power is New and Different
Certain aspects of corporate discursive power ‘of old’ and this ‘Modern Bigness’ discursive power overlap. Discursive power of Big Tech platforms by way of direct lobbying, funding academic research or think tanks, or funding political campaigns is seemingly no different. However, when digital informational channels of the same conglomerate can be used to amplify their own/chosen other voices, or strengthen a campaign, and where user tracking can be mined for the granular data about user behaviour and highly individualized targeting of audiences (Morozovaite & Gerbrandy 2024, forthcoming), Big Tech companies gain a novel edge. Their political and public lobbying efforts can shift the Overton window; the specific lobbying surrounding possible AI regulation can serve as a case in point. A further difference lies in the comingling of activities, where some Big Tech platform companies are also media companies. This influences the traditional media landscape, both as to its structure, its financing, and to the content provided. Big Tech companies can be both providers of media content, curators of media content, media producers, and have impact on the plurality of media voices. They are, in essence, self-regulators of access to the public domain discourse.
Discursive Power and Competition Law
Are these problems an issue for competition law, though? Competition law is aimed at redressing negative effects that stem from an abuse of a dominant market position. This means that for enforcement one has to establish such a dominant market position and have a credible ‘theory of harm’, a notion of why the behaviour of the dominant company creates a competition law relevant harm.
In current European competition law establishing a ‘dominant position’ focuses on the market position in a ‘relevant market’. This means finding where competitive pressure exists (or potentially arises) and examining the market power of the company. For some of the examples of discursive power mentioned above such an exercise seems doable: a market for (specific) media services might be delineated, from which a position of ‘power’ over media content might follow. Or a market for (targeted, political) online advertising. However, when considering other examples such as content moderation practices this becomes more difficult, just as with political lobbying, bolstered by influencing and shifting the Overton window on a certain topic. And it becomes almost impossible when considering the landscape of digital public discourse. What would be the relevant ‘market’ here? One may be able to carefully delineate markets, its players, and their respective positions but in the current digital media and information landscape the boundaries between political, commercial and social ‘markets’ are increasingly blurred. Also, separating ‘markets’ misses the point of the combined construct of discursive power. To grasp ‘dominance’ in the sense of discursive power means a different way of constructing ‘power’ in this digital world.
A similar issue arises when considering theories of harm. In traditional competition law, harms that are competition-law-relevant are generally related to consumer welfare: a price increase, a quality loss, a degradation of service-levels, a brake on innovation or choice. Some of the harms related to the exercise of discursive power will fit this frame, such as limiting consumer choice by lock-in mechanisms, limiting access to platform or other content, lowering the quality of service by either aggressive or too lax content moderation practices. Competition law can be a tool for combatting these harms. However, in traditional interpretations of competition law, non-market values – such as protecting democracy – are difficult to incorporate, though there are suggestions in current literature proposing to widen harm-theories to a ‘citizen welfare’ idea, making competition law better suited to address discursive power.
The Limited Utility of a Competition Law- Based Approach to Regulation
The above means that, both in theory and in practice, some instances in which discursive power is wielded by dominant companies in a way that leads to negative effects on consumer welfare (broadly construed), or negative effects on citizen-welfare, might be countered by competition law (if dominance is established).
This way of analysing instances of discursive power under an abuse of dominance regime may, however, overlook the essence of Modern Bigness’ discursive power and does not capture the possible broader harms. Firstly, this way of isolating ‘negative effects’ (on the market, on consumer welfare, or even – in a broader notion – on citizen welfare) misses that it might be a combination of harms that is the point of concern. Secondly, it misses the point that even combined effects do not stem from a position of isolated power, but of Modern Bigness power: a composite of power with different dimensions. Thirdly, an instance-based focus might overlook that the negative impacts may stem from the fact that it is mostly a handful of Big Tech corporations, next to some ‘old media’ companies, which together shape the direction of the public (online) discourse. Fourthly, such a starting point might overlook that it is also these same companies that, to a certain extent, have power over (technological) innovations, including those that immediately impact the public digital sphere. And, fifthly and finally, such a case-by-case analysis might miss the long-term effects, for example on the robust functioning of democratic societies.
What Else is There?
An open public digital discourse with a plurality of voices is of utmost importance for a vibrant democratic society. Such a democratic society (based on the rule of law) to us is a society that is something to strive for, to cherish and to protect and competition law may have a role to play here. IN light of this normative setting, there are different ways to analyse the negative effects of discursive power of big tech corporations in light of the Modern Bigness concept: focus on specific harms or focus on combined effects; focus on competition-relevant harms, which might be more or less broadly construed, or focus on (combined) effects on an open (digital) public discourse and well-functioning of democracies. It is only then that the long-term impacts really come into view. Competition law, with its focuses on market power and combating negative effects on a case-by-case basis, on quality, quantity, price, choice or innovation, can be a relevant instrument, if it fits existing or novel theories of harm.
However, there is a limit to what competition law can do.
It cannot readily deal with negative effects that are too far removed from the competition law accepted theories of harm. Unless, of course, one argues for an ever wider set of relevant harms, which might include protection of democracy, protection of media plurality and voices, and protection of citizen welfare and wellbeing more generally. Though we support that there are arguments in the EU Treaties’ set-up and its constitutional values to wield competition law in a broader way, there does come a point where its limits are reached.
We seem to run into those limits especially when it comes to the combined, long-term, negative effects on the public digital discourse and thus democracy, that can stem from wielding discursive power without indemnity. Other regulatory instruments whether existing ones such as the DMA and the DSA or novel ones have a complementary role to play. But when it comes to counter these long-term negative impacts of Modern Bigness’ discursive power on the digital public sphere and democracy, perhaps we need to shift our focus, either by creating counterpower(s) or by looking at limiting Modern Bigness’ power itself.