Non-State Actors and the Global Governance of Artificial Intelligence

What is the proper role of non-state actors in the rapidly growing global governance of artificial intelligence? There is by now a host of regulations, rules and policies to guide the development and deployment of AI globally, with the aim of solving problems and providing goods associated with AI technologies. Although the coming EU AI act represents a shift toward ‘hard law’, a large share of the global governance of AI still consists of ‘soft law’ initiatives authored by or involving non-state actors. Non-state actors are often assumed to play a central the role in the democratization of global governance. The inclusion of international non-governmental organizations (INGOs), advocacy groups and social movements representing civil society is usually thought to help address a perceived democratic deficit in global governance generally, by promoting increased representation, transparency and openness towards citizens around the world. Unlike many other policy areas in global governance, however, the non-state actors engaged in AI governance include not only civil society actors but also the very same multinational tech companies that develop the AI technology that is to be regulated. These companies possess large economic resources, power, and expertise concerning AI technology, and how they govern their platforms has repercussions for the democratic character of society at large. We cannot simply assume that such non-state actors have a democratizing effect, but it would also be a mistake to disregard their potentially beneficial influence on the global governance of AI, given their unique position at the head of the AI transformation. Surprisingly, the task of theorizing the normative requirements of these actors from a democratic point of view has been largely ignored. To fill this void, this paper analyzes the democratic credentials of non-state actors in global AI governance. I adopt an ecumenical understanding of democracy, assuming that governance arrangements are democratic to the extent that they derive their authority from the people and that those affected by the decisions have a direct or indirect say in the decision-making. This paper suggests that on a plausible account of legitimate AI governance, non-state actors are not authorized to wield public authority as ‘democratic agents’. Yet, to the extent that they are justified in exercising what I call moral, epistemic, or market authority, non-state actors can nevertheless contribute to the strengthening of the prerequisites for the democratization of the global governance of AI, as ‘agents of democracy’. On this account, non-state actors wielding moral authority, such as advocacy groups, social movements and INGOs, do not contribute to the democratization of global AI governance, even if their morally desirable goals typically strengthen many values associated with a well-functioning democracy. On the other hand, non-state actors including powerful tech companies can in principle be said to have democratic credentials, if their efforts as agents of democracy strengthen the prerequisites for democratization.