OpenAI Safety Leaders Exit as Company Dissolves Key Alignment Team
OpenAI's head of alignment resigns and criticizes the company for prioritizing products over safety. The departure follows the exit of chief scientist Ilya Sutskever and the dissolution of their 'Superalignment' team.
Jan Leike, the head of alignment at OpenAI, resigns from the company and publicly criticizes its current direction. He states that safety culture and processes take a backseat to shiny products, revealing that he disagrees with leadership about core priorities until reaching a breaking point. He warns that the company shoulders an enormous responsibility on behalf of humanity and needs to get incredibly serious about the implications of artificial general intelligence.
This high-profile departure happens just days after chief scientist Ilya Sutskever announces his exit from OpenAI. Both leaders co-lead the 'Superalignment' team, which focuses on ensuring advanced AI systems remain safe and beneficial. Their resignations follow months of internal tension surrounding the brief ousting of CEO Sam Altman in November 2023, an event that highlights ongoing disputes about the safety of the company's technology.
Bloomberg reports that OpenAI officially dissolves the 'Superalignment' team and folds it into broader research efforts across the company. Before leaving, Leike addresses OpenAI employees directly, urging them to feel the gravity of building artificial general intelligence and to ship the cultural change that is needed within the organization.