School of Social and Political Science

Description

AI safety – the goal to create ‘provably beneficial’ AI systems – has recently gained momentum. Join Dr Apolline Taillandier to examine the construction of AI safety in the USA and the UK from the 2010s onwards, drawing on interviews with AI safety experts and qualitative analysis of philanthropic, machine learning and policy discourse.

Abstract

Artificial Intelligence (AI) safety – the goal to create ‘provably beneficial’ AI systems – recently gained momentum, moving from computer science, AI ethics, and debates in effective altruism to the forefront of national and global technology policy. This paper examines the construction of AI safety in the United States and the United Kingdom from the 2010s onwards, drawing on interviews with AI safety experts and qualitative analysis of philanthropic, machine learning and policy discourse and instruments. Tracing how AI safety experts engage in boundary-work to redefine both AI research and ethics, and how they produce knowledge and ignorance about AI progress, including through predictions and forecasts of future machine learning capabilities or computing power, I argue that entrepreneurial insight and moral arguments play a central role in the management of AI risk and the scientific, moral, and economic valuation of AI safety. The paper brings together the sociology of philanthropy, the sociology of knowledge, and social studies of AI and ethics, shedding light on the worldmaking power of AI expertise.

Speaker Bio

Dr Apolline Taillandier is a political theorist and historian of political thought specialising in the history of liberalism, feminism, and technology from the mid-twentieth century onwards. At present, she is a Postdoctoral Research Fellow at the Leverhulme Centre for the Future of Intelligence, the Department of Politics and International Studies and Newnham College at the University of Cambridge jointly with the Center for Science and Thought at the University of Bonn, and a visiting researcher at the School of Social and Political Science at the University of Edinburgh. Previously, she was a Doctoral Fellow at the Max Planck Sciences Po Center in Paris, and she held visiting fellowships at the University of California, Berkeley, and Cambridge. She is currently completing a monograph entitled In the Name of Posthumanity on the history of postwar transhumanism in the United States and Britain, and her new project traces the circulation of feminist ideas in computer education in the US, the UK, and France.