Artificial intelligence: promises, perils—and political economy

The rapid rise of Artificial Intelligence (AI) has prompted policymakers and political scientists alike to evaluate the (mis)uses of this technology for business and politics. On the one hand, political scientists are concerned that granting the state control over AI could render it an ultra-powerful agent for surveillance. On the other, granting the market control over AI might allow particular firms to displace states as sites of authority. A large body of work on “AI Governance” has hence burgeoned. What remains under analyzed are the political determinants under which states or markets may accumulate said power and what regulatory regimes might subsequently result. This article argues for introducing a macro-political economy approach to explain the perils of AI and to then craft equitable solutions. The construction of AI regulatory frameworks must be guided by how workers and firm preferences aggregate, as well as under what institutional constraints. We begin by giving an overview of existing literature. We then turn to the framework of comparative political economy to identify macro-level variations in AI development and regulation across countries. Finally, we emphasize how, unlike previous technologies of automation, AI impacts higher-income, non-routine workers and how this politics shapes contemporary discussions about AI governance.