Skip to content

Aytac 2024 – Big Tech, Algorithmic Power, and Democratic Control

In brief

Aytac argues that Citizen Boards of Governance (CBGs) – bodies of randomly selected citizens authorised to govern the algorithmic infrastructure of Big Tech platforms – are the best way to democratically contain Big Tech's algorithmic power in the digital public sphere (p. 1431) [U]. The main advantage is that CBGs tackle the concentrated power of tech corporations without giving too much power to governments (p. 1431) [U]. He shows why this is better than ordinary state regulation or market mechanisms (pp. 1431, 1436–1437) [U]. The underlying commitment is to keep the public sphere autonomous from centres of power, whether corporations or governments (p. 1431) [U].

Research question and main claim

  • Two arguments: first, Big Tech's concentrated powers give distinctive reasons to democratise algorithms in the digital public sphere; second, these problems can only be addressed via a democratic route beyond government regulation or market mechanisms (p. 1431) [U].
  • Desideratum: ‘The functioning of the public sphere should not be controlled by any group of political, social, and/or economic elites’ (p. 1433) [U].
  • Scope is limited to sorting and recommendation algorithms of the largest platforms and search engines (Facebook, YouTube, Twitter, Google) (p. 1431) [U].

Theoretical perspective

  • Democratic theory: deliberative (Habermas, Christiano), republican (Pettit) and realist/plebeian approaches (Bagg, Klein, McCormick) (p. 1433) [U].
  • Literature on democratic innovations and mini-publics (Beauvais and Warren, Landemore, Lafont) (p. 1438) [U].
  • Democratic theory of the business corporation (Ciepley, Bennett and Claassen) (pp. 1432, 1438) [U].
  • Bachrach and Baratz's ‘second face of power’ (p. 1436) [U]. Pamuk on adversarial proceedings (p. 1441) [U].

Empirical material and context

  • No empirical study of his own. Draws on research on algorithmic amplification on Twitter, news media dependence on platforms and similar (pp. 1434–1436) [U].
  • The EU's Digital Services Act is mentioned as an example of existing regulation (p. 1437) [U].

Method

  • Normative political theory: diagnosis of a problem, an institutional proposal, and replies to objections (pp. 1432, 1442–1443) [U].

Key concepts

Concept How the source defines or uses it Page U/T
corporate algorithmic power Big Tech's automated capacity to shape user experience, dissemination of information and online visibility 1433 U
digital public sphere Social media, search engines and other technologies that disseminate information on a mass scale 1431 U
public sphere desideratum No group of elites should control the processes of public opinion formation 1433 U
Citizen Boards of Governance (CBGs) Randomly selected citizen bodies integrated into Big Tech's corporate governance, authorised to oversee and shape algorithms 1431–1432, 1438 U
adversarial proceedings Confrontational debates where expert claims are tested against opposing experts 1441 U
quasi-output legitimacy Legitimacy from a body's positive effect on the overall balance of power 1443 U

How privacy and private data are understood

  • Not the focus. Aytac explicitly chooses algorithms rather than the governance of user data as his topic (p. 1432) [U].
  • Users' personal data, harvested by Big Tech, give users an extra set of interests to be protected on grounds of ownership, privacy and anonymity; these can be represented through separate mechanisms (p. 1438) [U].
  • Data protection is mentioned as one of the areas states already regulate (p. 1437) [U].

Arguments and findings

  • Big Tech's algorithms give a small group of corporate elites significant control over online visibility, which raises questions of political equality (pp. 1434–1435) [U].
  • Big Tech has become a centralised gatekeeper in the news media industry, constraining journalistic autonomy (p. 1435) [U].
  • The lack of democratic control deprives citizens of a tool to counteract elites' control over attention and agenda setting (p. 1436) [U].
  • Breaking up Big Tech or creating markets for algorithms would fragment the shared public sphere (pp. 1436–1437) [U].
  • Extensive state regulation risks jeopardising the public sphere's autonomy as much as Big Tech does; politicians face self-interested pressures (p. 1437) [U].
  • CBGs: separate boards of 51 citizens per company, three-year terms, stratified random sampling, an expert committee appointed by political parties, and powers to oversee, veto and commission algorithms (pp. 1440–1442) [U].

Limitations the source itself mentions

  • The proposal is ‘more utopian’ than Social Media Councils (p. 1432) [U].
  • He provides only a general outline, not a legislative proposal; details should come from a democratic process (p. 1440) [U].
  • Citizen members lack technical skills to design algorithms (pp. 1441–1442) [U].
  • Objections on expertise, shareholder primacy and legitimacy are discussed (pp. 1442–1443) [U].

What the source does not cover

  • Norway or Nordic regulation [T].
  • The state's own use of data [T].

Possible relevance for the thesis

  • Represents a position sceptical of giving the state more power over digital infrastructure, which can be compared with Norwegian documents that give the state a central regulatory role [T].
  • Explicitly separates algorithmic power over the public sphere from questions of personal data (pp. 1432, 1438), which may help clarify what the thesis covers [T].

Connections to other sources

Key quotes

“The main advantage of CBGs is to tackle the concentrated powers of private tech corporations without giving too much power to governments.” (p. 1431)

“The functioning of the public sphere should not be controlled by any group of political, social, and/or economic elites.” (p. 1433)

“A major problem with this solution is that extensive state involvement might jeopardize the autonomy of the digital public sphere as much as Big Tech’s concentrated powers.” (p. 1437)

Open questions for Rebekka

  • Is the digital public sphere part of your research question, or is this source mainly background on democracy and platforms?
  • Do Norwegian documents see the state as the natural actor to regulate tech companies? How does that compare with Aytac's scepticism?
  • Is the distinction between algorithmic power and personal data useful for delimiting your thesis?