Tantum Collins on what he’s learned as an AI policy insider at the White House, DeepMind and elsewhere

Abstract

The interviewee, a former White House official and researcher at DeepMind, discusses the increasing prominence of AI in government policymaking and the risks of AI-powered autocracy. The risks of both misuse and misalignment of AI are being taken more seriously in the US government. Though the US is considered a leader in AI research and development, Chinese AI research output is now greater than in the US, and China is rapidly catching up. There are significant differences in the way the US and China approach AI regulation. The interviewee believes that more public funding and international coordination of AI research are necessary, and highlights the need for a clear benchmarking and standards regime to measure the capabilities of AI models and inform user decisions. The interviewee discusses the tensions between the AI safety community and those focused on AI ethics, suggesting that both communities have legitimate concerns that need to be addressed. The interviewee also discusses the potential for AI to transform the labour market and whether AI policy will become a partisan political issue in the US. Finally, the interviewee explores the implications of panpsychism for AI systems and the potential for AI to reshape the very concept of democracy. – AI-generated abstract