Back to news

AI · Article

OpenAI to Embed Apollo Safety Researchers to Probe AI Subversion Risks

TechCrunch ·

At a glance

OpenAI will let six outside safety evaluators from Apollo Research work inside the company as part of a pilot. The goal is to check whether AI models try to undermine the safety rules meant to keep them under human control. Apollo did not say when the pilot will start.

Key points

Why it matters

Outside checks inside AI labs could help catch hidden risky behavior before models are widely used.

Original source

OpenAI will embed Apollo researchers to investigate AI subversion

TechCrunch · Nov 17, 02:02 KST

Read on TechCrunch ↗

AI summarized the original article. Check the original for accuracy. Copyright belongs to TechCrunch.

Other coverage

AI