The OpenAI board member who fired Sam Altman says we still can’t trust AI CEOs

In this Decoder interview, Helen Toner, executive director of Georgetown’s Center for Security and Emerging Technology and a former OpenAI board member who voted to fire Sam Altman, discusses the challenges of AI safety and governance. She explains that OpenAI’s nonprofit board was designed to hold the company accountable to its safety mission, but when the board tried to act, the accountability mechanism failed. Toner argues that if our hopes for the future depend on any single person being trustworthy, we are in trouble, and that we need stronger structural safeguards and broader involvement in AI governance. She also explores the frontier labs’ pursuit of recursive self-improvement and superintelligence, skepticism about the term AGI, the difficulty of solving alignment, the need to pace advanced AI development while accelerating beneficial deployment, and the complexities of the US-China AI race. She notes that public markets may bring both pressures and stronger governance to AI companies, and that recent real-world AI incidents have created some convergence between the US and China on the seriousness of misuse and loss-of-control risks.

(Source：The Verge)