FISHBOWL Session - AI for Evidence-Informed Policymaking: Developing a Framework
A new generation of AI-based tools could present an opportunity in the near future to dramatically improve science advice, making it more agile, rigorous, and targeted. In the future, new AI-based platforms should be able to make evidence syntheses less time-intensive and free subject matter experts to focus on more complex analytical aspects of the process. However, leveraging such tools for good will require science advisers and policy institutions to create guidelines and carefully consider the design and responsible use of the nascent technology. And collaboration will be needed to build new tools responsibly. The technical know-how will likely come from academia and technology companies, while demands for robust governance, transparency, and accountability can only be met by governments. Science advice needs to be scientifically credible, politically legitimate, and relevant to the needs of policymakers. And that must remain so if AI tools are used, which has consequences ranging from appropriate system design to content selection, and governance.
In this session, the panel members will present a series of proposals for how to harness responsible AI to support government science advice and then invite session participants to share their perspectives and participate in developing the framework
Panelists:
Chris Tyler: Associate Professor in Science Policy and Knowledge Infrastructure, Department of Science, Technology, Engineering and Public Policy (STEaPP), UCL – UK
Rhona Mijumbi: Co-Director, The Center for Rapid Evidence Synthesis (ACRES); Africa Evidence Network - Uganda
David Budtz Pedersen: Professor, Department of Communication & Psychology, Aalborg University – Denmark
Ronald Munatsi, Executive Director, Zimbabwe Evidence Informed Policy Network – Zimbabwe