The Global Governance of Artificial Intelligence and its Democratic Challenges

The impact of Artificial Intelligence (AI) on society raises inherently normative questions, which call for political answers.

In democratic societies, the goals and means of AI governance should be identified and decided upon collectively by those who are affected by AI technology, yet what this entails is currently undertheorized.

With almost 500 AI governance initiatives launched in the last five years, this four-year research project allows two political scientists and one philosopher to redress this.

Assuming that a key property of good governance is democratic legitimacy, and that a study of AI governance should primarily focus on the global level, the project’s aim is twofold: (1) to develop a democratic theory of global AI governance, and (2) to apply this theory to assess the democratic quality of existing instances of global AI governance.

The overarching research question is: under what conditions is global AI governance democratic?

Relying on methods like conceptual and normative analysis, and a coherentist account of justification, the project answers this question through three work packages: (1) a suitable theoretical framework is developed (months 1-12); (2) this framework is drawn upon to develop a democratic theory of global AI governance (months 13-34); (3) which is then applied to instances of global AI governance (months 35-48).

A general evaluative typology is developed, mapping out the democratic demands on existing entities and their decision-making.