Team UK

Debates in AI research can be expected to exert an important influence on the formation of public discourse about AI. In the last years, knowledge claims about the unprecedented, revolutionary, capacities of machine learning systems to detect patterns in data at scale have been invoked as a key driver of change across science, policy and media debates. Many of the AI controversies that have occurred across these spheres centre on the impact of science and innovation – from facial recognition to autonomous vehicles – on society. But what about controversies within and about AI research itself? What concerns have recent developments in AI raised among research communities in the UK, Germany, France and Canada? Are the issues identified by computer scientists the same or different to those raised by social science and humanities scholars? And to what extent do wider communities of expertise – policy-makers, legal experts, ethicists, activists and citizens concerned with AI – engage with AI knowledge claims?

The UK team for Shaping AI will map controversies in and about AI research across the sciences, humanities and extended peer communities (Funtowicz and Ravetz, 1993). Combining digital methods of issue mapping (social media & Web analysis, scientometrics) with ethnographic engagement with AI research, we will use network and textual analysis to identify actors, objects and concerns animating controversies in AI research in the relevant period (Marres, 2015). In doing so, our aim is to develop a contextual understanding of the knowledge propositions that organise – and/or fail to organise – controversies in and about AI research across different fields of expertise, and to contribute to Shaping AI’s wider investigation as to whether distinctive modes of problematization are detectable in AI controversies (Roberge & Castelle, 2021).