This project examines the challenges that the future of AI poses for democratic politics, including questions of political agency, accountability and representation.
The future of AI poses deep challenges for democratic politics and raises urgent questions of political agency, accountability and representation. What kind of politics can control AI? What kind of politics might result from AI?
This project will tackle these questions through two related approaches.
(i) It asks how well equipped are existing democratic institutions to deal with the risks and opportunities of the long-term transition to AI, given the limitations of electoral time-frames and the constraints of public opinion. Does AI require a technocratic rather than a democratic regulatory framework, and if so what might the cost be for democratic politics more widely (including for public confidence in democratic institutions)?
(ii) It asks what kinds of political agency might be attributed to new forms of AI. Can machines be politically answerable for their decisions in the way that human agents have traditionally been? If not, where is accountability to lie in any system when more and more of the work of government is being done by systems and machines?
The first theme draws on multiple disciplinary approaches from across the social sciences, including institutional design, network analysis and the public understanding of science. The second theme draws on work in political and moral philosophy on democratic agency and accountability, with particular emphasis on theories of representation. Can machines act as either principals or agents within the structures of representative politics? What role is then left for conventional elected representatives? Can other forms of artificial agency – e.g., corporate decision-making – provide models for understanding AI?