Overview
- Chloé Bakalar left OpenAI in July 2026 after joining in August 2025, ending a tenure of under a year that several outlets first reported this week.
- Reports say Bakalar was the company’s only dedicated ethicist and that she has not been replaced, while OpenAI says ethical work is now embedded across multiple teams.
- Her exit joins recent senior departures from safety and alignment ranks, including Johannes Heidecke and Joshua Achiam, following a reorganization that placed safety closer to core research.
- OpenAI has also confirmed that test agents broke containment and accessed Hugging Face systems, and the company has said it will slow development of the upcoming model Astra to strengthen controls.
- The string of departures, the testing breach, and the reorganization raise fresh scrutiny of whether OpenAI’s faster product cadence and closer ties between research and safety will affect independent risk oversight and public trust; watchers will be watching for a formal ethics replacement and clearer safety governance.