Three fired OpenAI safety researchers challenge leak allegations
In brief
- Three OpenAI safety researchers fired for alleged leaks dispute the accusations.
- OpenAI never specified in writing which data they supposedly mishandled, they claim.
- Their case raises questions about AI model transparency and internal security practices.
Three former OpenAI safety researchers—Tomek Korbak, Mikita Balesni, and Jasmine Wang—are pushing back against their dismissal for allegedly leaking confidential information. They claim OpenAI never specified in writing which data they supposedly mishandled.
The researchers, known for advocating more transparent AI systems, argue their termination could deter independent auditors from collaborating with tech firms. OpenAI stated the dismissals were unrelated to their safety concerns but did not detail the alleged policy violations.
Their case highlights a technical and strategic issue: external access to AI models' reasoning processes. The researchers recently co-authored a paper calling to preserve this capability as companies reduce model transparency.
What it changes for you
- Public SectorThe firings could discourage independent auditors from working with tech companies, limiting external oversight of AI systems.
This summary gives you the essentials. The full article, its figures and quotes are on The Decoder.
Read the original articleAI-generated summary of the source, checked by rules. How we write · Report an error