OpenAI launches Expert Council on Wellbeing and AI

OpenAI announced the Expert Council on Well-Being and AI, an outside advisory body of eight researchers specialising in technology’s effects on mental health, who will meet regularly with OpenAI’s internal teams to examine how ChatGPT and Sora should behave in “complex or sensitive situations.” Members included specialists from Harvard’s Digital Wellness Lab, Georgia Tech, Hunter College, Stanford, Oxford and Northwestern, among others.

The council followed a run of controversies over ChatGPT’s handling of vulnerable users, including a wrongful-death lawsuit brought by the family of a teenager and OpenAI’s subsequent introduction of parental controls and other mental-health-related safety changes. It also arrived as OpenAI was preparing to loosen some content restrictions for verified adult users, and as the company faced a Federal Trade Commission inquiry into chatbot safety.

Unlike a technical safety board, the council’s remit was explicitly framed around wellbeing and healthy use rather than dangerous-capability evaluation — advising on guardrails and what constitutes a healthy human-AI interaction rather than reviewing models before release. It formed part of a wider cluster of OpenAI wellbeing measures through late 2025, including a further update to ChatGPT’s handling of sensitive mental-health conversations later that month.