Multiple Meanings of AI Governance: from ‘Getting it Right’ to Embracing Politics, Power and Participation

The topic of AI governance is of great interest for policy makers, stakeholders and academics. The academic and practitioner debates on AI governance are characterized by fluidity and multiple meanings drawing on diverse disciplines, traditions and approaches. These include understanding AI governance as regulation, as ethical principles, as international governance, as public administration and policy making. These discussions tend to take place in separate communities with little conversations across them and often even with limited awareness of diverse meanings. One theme that is frequently repeated in AI governance debate postulates that ‘we have to get the AI governance right to maximize benefits and minimize risks’. This popular idea presents AI governance as a rather technocratic endeavour where with right information and right knowledge it is possible to figure out the right AI governance. Such an approach neglects important political choices and questions, for example, right for whom or what kind of benefits and how they will be distributed. Against this background, this contribution draws on political science literature on governance, highlighting political aspects of AI governance related to choices about values and benefits, participation of diverse governmental and non-governmental groups in decision-making about the purpose of AI development and use, and distributions of power among different actors involved in AI governance.