OpenAI Fires Three Safety Researchers Over Unauthorized Data Leak to External AI Watchdogs.

OpenAI Terminates Three Safety Researchers for Unauthorized Internal Data Leak to External Safety Organizations

OpenAI has officially confirmed the termination of three research personnel within its artificial intelligence safety division following an internal investigation. The company stated that the employees were let go after being found in violation of corporate policy and operational protocols by sharing confidential internal data with external AI safety evaluation entities.

Corporate Policy Violation, Investigative Details, and Institutional Ties

• Unauthorized Disclosure & Protocol Breach:

  • Breach of Confidentiality: OpenAI determined that the three researchers bypassed internal disclosure channels and compliance protocols to transfer non-public research and safety evaluation data to third-party safety organizations.

  • Anonymity of Parties: OpenAI has declined to publicly disclose the names of the terminated researchers or identify the specific external AI safety organizations that received the unauthorized materials.

• Reporting Links to AI Safety Institute METR:

  • Industry Reporting: Reports from The Wall Street Journal indicate that at least one of the terminated researchers currently serves as an advisor to METR (Model Evaluation and Threat Research), a prominent non-profit AI safety research organization.

  • Existing Institutional Collaboration: METR has previously partnered with OpenAI, Anthropic, and other frontier AI developers to conduct third-party red-teaming and safety benchmarks on flagship AI models prior to deployment.

 

 

Source: The Wall Street Journal 

💬 AI Content Assistant

Ask me anything about this article. No data is stored for your question.

OpenAI Terminates Three Safety Researchers for Unauthorized Internal Data Leak to External Safety Organizations

OpenAI has officially confirmed the termination of three research personnel within its artificial intelligence safety division following an internal investigation. The company stated that the employees were let go after being found in violation of corporate policy and operational protocols by sharing confidential internal data with external AI safety evaluation entities.

Corporate Policy Violation, Investigative Details, and Institutional Ties

• Unauthorized Disclosure & Protocol Breach:

  • Breach of Confidentiality: OpenAI determined that the three researchers bypassed internal disclosure channels and compliance protocols to transfer non-public research and safety evaluation data to third-party safety organizations.

  • Anonymity of Parties: OpenAI has declined to publicly disclose the names of the terminated researchers or identify the specific external AI safety organizations that received the unauthorized materials.

• Reporting Links to AI Safety Institute METR:

  • Industry Reporting: Reports from The Wall Street Journal indicate that at least one of the terminated researchers currently serves as an advisor to METR (Model Evaluation and Threat Research), a prominent non-profit AI safety research organization.

  • Existing Institutional Collaboration: METR has previously partnered with OpenAI, Anthropic, and other frontier AI developers to conduct third-party red-teaming and safety benchmarks on flagship AI models prior to deployment.

 

 

Source: The Wall Street Journal