OpenAI has terminated its relationship with three researchers following an internal investigation. The company stated that these individuals violated established protocols regarding the access and handling of sensitive information, which undermined the trust required for their work.
The former employees have been identified as Jasmine Wang, Tomek Korbak, and Mykyta Baliesnyi. According to reports, the individuals were accused of sharing confidential company data with a third-party AI safety organization without authorization.
Tomek Korbak was a member of the OpenAI safety team and served as a technical contact during investigations into the Hugging Face incident. Jasmine Wang and Mykyta Baliesnyi were involved in model alignment research, focusing on ensuring AI models adhere to human intent.
OpenAI maintains that it supports independent safety evaluations and provides external researchers with access to training and deployment data. The company emphasizes that such assessments are necessary to identify potential risks and verify the effectiveness of existing security measures.
