The roles of government and power asymmetries in the governance of generative AI

The governance and policy-making of Artificial Intelligence (AI) is often characterized by huge power asymmetries, where power is seen as highly concentrated in a small number of big tech companies, while society is perceived as having rather limited role to play. In this context, it is of particular importance to interrogate the roles of government in re-shaping or re-enforcing existing power asymmetries. Here it is productive to think about the various roles of government in line with the distinction made by Sartori (1970) highlighting that it is important to consider not only the ‘difference in degree’ but also ‘difference in kind’. Namely, when analysing the roles of government in governing AI, it is illuminating to think not only about more or less government involvement but also about different types of its involvement, such as regulator, risk manager, funder, buyer or enabler of public participation. This contribution will examine various roles of government in governance of generative AI by analysing recent policy documents. It will reflect on the implications of these role on existing power asymmetries in AI: is government reinforcing power of big tech companies or reshaping power relations by facilitating involvement of diverse groups and distributing power more widely and equally?