IT之家 · ITHome
Three fired OpenAI safety researchers reject leak allegations
Three safety researchers dismissed by OpenAI deny leaking confidential information and suspect their advocacy for AI transparency contributed to their firing. OpenAI says an internal investigation found breaches of sensitive-information rules and that safety concerns were not the reason for dismissal. The researchers say they received no written explanation identifying what information was allegedly mishandled and have urged the board to strengthen external safety partnerships.
What to know
- 01
Tomek Korbak, Mikita Balesni and Jasmine Wang coauthored a paper advocating preservation of the ability to examine models’ intermediate reasoning steps.
- 02
OpenAI says Signal is not an approved communication tool, but emphasizes that this was not the reason for the dismissals.
Why it matters
The case concerns AI confidentiality practices and external safety collaboration, with the company and researchers offering conflicting accounts of the reasons for dismissal.
AI-assisted brief. Read the source for full details. Editorial & sourcing policy
