SAN FRANCISCO — Chloé Bakalar, OpenAI's AI ethics lead, left the company in mid-2026 after less than one year in the role. She joined in August 2025, making her exit the latest in a series of personnel changes across the company's safety and ethics divisions.
The departure continues a pattern that stretches back to 2024: roughly 50 percent of OpenAI's safety researchers have left since 2023, many citing what they described as a deprioritization of safety work as the company accelerated its product releases and commercial push.
Johannes Heidecke, who led Safety Systems, also left in July 2026. Lilian Weng, a former head of Safety Systems, departed around the same time. The turnover has hit the Superalignment team and other safety leadership roles hardest.
Neither Bakalar nor OpenAI has commented publicly on the reasons for her departure. The company has restructured during this period, folding safety efforts under its core research leadership.
OpenAI began as a nonprofit research lab and has since become one of the most valuable private technology companies in the world — a shift that introduced real commercial pressure. The original mission of developing AI safely for humanity's benefit now competes with product cycles and market demands.
The exits have a competitive dimension. Anthropic was founded by former OpenAI employees who left partly over safety concerns and has built responsible AI development into its core brand and market positioning.
OpenAI's ability to retain safety and ethics talent is a direct test of whether its stated commitment to responsible AI development can hold alongside its commercial expansion.
