The EU and US are taking different regulatory approaches to address the risks associated with AI, with global implications. However, little empirical evidence exists on public attitudes towards different types of AI regulatory approaches, and what factors could shift their choice of regime. This dissertation presents evidence from 1375 respondents across structured interviews and randomized controlled survey experiments with policymakers, their advisors, tech regulators, lobbyists, academics, and the general public across the EU and the US to investigate the factors that influence support for enacting and revisiting AI regulation. Sixty-five semi-structured interviews were conducted with AI experts and government officials to investigate factors that influence support for AI regulation and inform the development of a survey experiment. One thousand three hundred and ten participants were then randomly assigned to one of three conditions describing different types of risks in a survey experiment for explication to test for risks that influence attitudes towards AI regulation including disinformation, job-loss and online child safety. The dissertation presents evidence that political affiliation and ecological consciousness are strong predictors of support for AI regulation. US respondents were happier with what their government had done in terms of tech regulation than EU respondents. US and EU views on who should regulate AI, how, and how often are also presented. These findings have implications on public support for adaptive versus static regulatory approaches and factors that influence support for regulating emerging technologies.
