Policy as infrastructure: Enacting artificial intelligence and making Europe

At least in pop culture, artificial intelligence (AI) has long been associated with issues of security and feelings ranging from dread and fear to awe. From Terminator's Skynet to killer bees in Black Mirror, AI has variously been presented as (existential) threat, salvation, or transcendent next stage for humanity. In the contemporary world, governments around the world have suggested in speeches, national strategies, and government spending that AI is playing a crucial role in international security now and in the future. In April 2021, for example, the European Commission became the first governmental entity to initiate a legislative process for governing AI. These policy processes, however, take place as AI remains an umbrella term for a field of technologies-in-the-making, which is constantly moving and whose definition continues to be difficult to pin down.

In exploring how AI is being enacted through a particular effort – that of the European Union – to legislate for this technological field as it develops, we conceive of the European AI policy as infrastructure-in-the-making which allows us to explore policy processes as networked and assembled, as processual rather than stable, and as affective. In doing so, we unpack how the ways of doing and thinking AI are being infrastructured through policy. Beyond this, we are also concerned with what else is being produced in and through this policy infrastructure. We argue that specific enactments of AI in this infrastructure are entangled with and reinforce particular notions of ‘Europe’.