Public Utility Governance for Transformative AI

Our project investigates whether and how transformative AI (TAI) can and should be regulated as a public utility, exploring the conditions that justify such an approach (e.g., natural monopoly characteristics). Using methodologies including literature reviews, case studies, and expert interviews, our project proposes specific policy interventions to enhance market fairness and public benefit from transformative AI. We argue that TAI resembles key characteristics of traditional utilities (electricity, telecommunications, water) and historical regulatory regimes. Therefore, our proposed policy interventions include enforcing equitable pricing, interoperability, and non-discriminatory practices; regulating anti-competitive behaviors such as predatory pricing and vertical integration; supporting public and commons-based AI infrastructure; and implementing external audits, transparency requirements, and ethical development incentives. Ultimately, this research contributes specific, actionable policy recommendations to the existing high-level discussions on governing transformative AI as a public utility.