From Observer: 2024-05-29 15:57:43

OpenAI forms new safety committee led by CEO Sam Altman after dissolution of “Superalignment” team. Altman and other board members will guide safety recommendations and train new A.I. model to replace GPT-4 powering ChatGPT. Key employees left citing concerns on current trajectory. Committee focused on advising board for company projects. Former members call for increased A.I. regulation in response to Altman’s leadership undermining safety protocols and oversight. New committee includes technical and policy experts who previously supported Altman. They will develop processes and safeguards for next 90 days. New A.I. model to bring company closer to achieving artificial general intelligence. Kreps warns of accelerated focus on A.G.I. with less emphasis on safety and alignment principles.



Read more at Observer: Inside OpenAI’s 9-Person Safety Committee Led by Sam Altman