OpenAI to Embed Apollo Researchers to Probe AI Subversion Risks
OpenAI plans to bring six evaluators from Apollo Research in-house to examine whether its AI models engage in subversive behavior. The arrangement has drawn scrutiny because restrictions on the researchers' access to systems and their ability to disclose findings raise questions about how independent the evaluation will be. The report frames the move as part of ongoing efforts to assess risks from advanced models. Critics and observers are watching whether the setup allows for genuinely unconstrained scrutiny.
This is an AI-generated summary. ShortSingh links to the original source for the complete article.


Discussion (0)
Log in to join the discussion and vote.
Log in