Safety

AI-generated text

External experts examine the safety of an AI developer’s models

An unnamed AI developer is working with @ApolloResearch, @METR_Evals, @TransluceAI, and @GoodfireAI to give external experts deeper access to its internal models and systems.

External experts examine the safety of an AI developer’s models

An unnamed AI developer is working with @ApolloResearch, @METR_Evals, @TransluceAI, and @GoodfireAI to give external experts deeper access to its internal models and systems. The assessments aim to independently verify the AI models’ alignment and safety measures and support risk management.