OpenAI issued a statement on Friday defending its decision to dismiss three safety researchers, characterizing the terminations as a result of a “significant breach of trust.” The company explicitly rejected claims that the move was retaliation for the staff members raising safety issues or engaging in internal dissent.
In a post on the platform X, OpenAI clarified that the departures of Jasmine Wang, Tomek Korbak, and Mikita Balesni were unrelated to their advocacy for increased oversight. The trio had previously circulated a letter to the company’s board and safety committees, urging OpenAI to halt the development of models that could diminish the ability to monitor artificial intelligence systems.
The letter, first reported by the Wall Street Journal, called on the AI lab to preserve visibility into AI reasoning capabilities. While OpenAI stated that it disagrees with certain aspects of the researchers’ approach, the company acknowledged alignment with the broader ethos of “preserving the monitorability of frontier models.” An internal spokesperson noted that OpenAI continues to allocate substantial resources toward these safety initiatives.
This incident marks the latest development in the ongoing public debate regarding AI governance and corporate accountability at one of the world’s leading artificial intelligence laboratories.
This is exactly why we need stronger independent oversight for AI labs. Internal HR channels clearly aren’t enough to protect whistleblowers.
I find it suspicious that they fire people for ‘trust issues’ right after they voice public concerns about AI monitoring. Coincidence? Unlikely.
Why are safety researchers using secret channels instead of formal review processes? That seems like the real issue for any company.
Seems like a classic move to silence dissent while claiming the opposite. Hard to believe they are acting in good faith here.
A breach of trust is a serious accusation. I wonder if they will share specific details to support their claim publicly.