OpenAI Says Departed Researchers Violated Sensitive Information Policies

OpenAI research leaders addressed the recent departure of Jasmine, Mikita, and Tomek, stating that an internal investigation found they violated clear policies on handling sensitive information. The company asserted that the investigation uncovered a significant breach of trust beyond what was outlined in the researchers' published letter, and they stand by the decision to end their employment.

Clarification on Safety Concerns

The statement explicitly denied that the terminations were related to raising safety concerns or speaking out. OpenAI emphasized that spirited and critical debates about safety and research are common and encouraged within the company, noting that they do not terminate employees for raising concerns. They described the outcome as deeply sad, acknowledging the researchers' contributions and willingness to challenge ideas.

Future Safety Assessments

OpenAI announced it is actively finalizing contracts with third-party safety assessors and plans to announce details in the coming weeks. The company reiterated its commitment to embedding external assessors and collaborating with independent safety organizations, stating that many researchers already work productively with these partners.

Monitorability Commitment

The company agreed with the researchers' letter that preserving the monitorability of frontier models requires an industry-wide commitment. OpenAI highlighted that monitorability has long been a core piece of its research program, citing investments in publications like Monitoring Monitorability, the open sourcing of monitorability evals, and the system card for GPT-6 Astra.