Governing AI – attempting to herd cats? Introduction to the special issue on the Governance of Artificial Intelligence

Artificial Intelligence raises new, distinct governance challenges, as well as familiar governance challenges in novel ways. The governance of AI, moreover, is not an issue of distant futures, it is well underway – and it has characteristics akin to ‘herding cats’ with a mind of their own. This essay introduces the contributions to the special issue, situating them in broader political and social science literatures. It then provides a sketch of an interdisciplinary research agenda. It highlights the limits of 'explainable AI', makes the case for considering AI ethics and AI governance simultaneously, identifies as an underappreciated risk 'system effects' that arise from the introduction of AI applications, and calls for policymakers to consider both the opportunities and the risks of AI. Focusing on the (ab)uses of AI, rather than the complex, rapidly changing and hard-to-predict technology as such, might provide a superior approach to governing AI.