OpenAI Safety Researcher Departs, Citing Unease Over Company's Direction

Compiled by the editorial desk with reference to official statements, public posts, and industry reports.

Rosie Campbell, a safety researcher who spent nearly three and a half years at OpenAI, has resigned, citing discomfort with the company's recent direction and the dissolution of its AGI Readiness team. In a personal Substack post, she shared a message she had posted on the company's internal Slack channel days earlier, explaining her decision.

Campbell's resignation follows closely on the heels of Miles Brundage's departure. Brundage, who served as OpenAI's AGI czar, left the company earlier this month, prompting the organization to disband the AGI Readiness team entirely. Campbell noted that Brundage's exit was a key factor in her own decision, stating that she believes she can now pursue the mission of ensuring safe and beneficial AGI more effectively from outside the company.

While Campbell did not specify the exact changes that troubled her, she referenced being unsettled by shifts over the past year. This period includes the failed attempt to oust CEO Sam Altman in November 2023, a move reportedly driven by board members who had concerns about the company's safety focus under his leadership. Altman was reinstated shortly after, but the episode highlighted internal tensions.

The past year has also seen several other notable departures from OpenAI as the company grapples with its growing commercial focus and rising valuation. Campbell's message expressed hope that the culture that made the company special to her would be strengthened rather than diminished. She also urged her former colleagues to remember that the mission is not simply to build AGI, but to ensure it benefits humanity.

Campbell's warning is stark: she hopes OpenAI will take seriously the possibility that its current safety approach may not be sufficient for the vastly more powerful systems that could arrive within this decade. Her concerns echo those of other former employees who have voiced similar apprehensions about the pace of AI development and the adequacy of safety measures.

OpenAI's trajectory has raised broader questions about the balance between rapid advancement and responsible deployment. The company's increasing commercial partnerships and product integrations have drawn scrutiny from safety advocates and industry observers alike. As more researchers leave, the internal debate over safety versus speed continues to unfold publicly.

A Pattern of Departures

Campbell is not alone in her concerns. Over the past year, several key figures have left OpenAI, citing similar worries about the company's direction. These departures have fueled discussions about whether the company's safety culture is keeping pace with its ambitions. The dissolution of the AGI Readiness team, in particular, has raised eyebrows, as it was seen as a critical component of the company's safety infrastructure.

In her message, Campbell expressed a sincere hope that the qualities that made OpenAI special to her would not be lost. Her departure serves as a reminder of the challenges facing AI companies as they scale, and the ongoing tension between innovation and caution.