B · Normal
[OpenAI introduces four independent agencies to carry out AI security assessment] November 18, local time, November 16, the US Open Artificial Intelligence Research Center (OpenAI) announced that it has signed an agreement with four independent assessment agencies to provide support for the security work of this artificial intelligence company. The four evaluation institutions are Apollo Research, METR, Transluce and Goodfire. They will initially conduct pilot evaluations focusing on model performance and security protection mechanisms. According to the agreement, OpenAI will open in-depth access to internal systems to the above-mentioned organizations, allowing external experts to review internal models, source code, training data and other information, and verify model alignment and security protection measures. (CCTV Finance)
Comments