The role of dissensus in policy making on AI

Over the past 6 years there have been substantial moves in governing AI, from the EU’s AI Act to Australia’s development of a national framework to guide the use of Generative AI in schools. Yet, despite this roll out of broader frameworks to govern AI in education what is lacking are the wide range of stakeholder views on using AI in education. This paper focuses on this area, premised on calls for increasing the types of expertise and experiences that are drawn upon to co-design policies on developing and using AI. The paper focuses on a key problem - of the risks of co-design and collaboration collapsing into furious consensus and becoming a process for the maintenance of existing power relations – especially that of global technology companies and governments – rather than creative governing possibilities.

The paper outlines work being undertaken as part of an international project on democratising policymaking about AI in education. We outline how we intend to disrupt consensus building through methods from work in technical democracy (Callon et al., 2009) that involves running workshops and building tools to build capabilities for people to create policy to guide the development and use of AI systems. The paper will focus on how our use these technical democracy approaches to develop collective policy making approaches is informed by political theories of dissensus (e.g., Ranciere, 2003), in which rather than consensus and co-option, the aim is for shared uncertainty and a more expansive, though provisional, policy making apparatus.