Anthropic pledges outside evaluators permanent access to its AI systems
Dario Amodei is calling for an AI industry slowdown with a three-part plan. He says outside evaluators would check Anthropic’s safety measures, report incidents and assess model alignment during training.
TLDR
On September 12, Dario Amodei announced an essay arguing that the AI industry should slow down. He said Anthropic was unilaterally committing to the first step of his three-part plan: giving third-party evaluators permanent, employee-level access to its systems to verify adherence to safety measures, report incidents and assess model alignment during training.
Days earlier, on September 9, a researcher announced their resignation from Anthropic, saying they had spent three years doing pretraining research at Anthropic and OpenAI. They accused both companies of acting irresponsibly by racing toward self-improving superintelligence and “gambling with our lives.”
Combined views
5.9M
4 Sources, first seen ago
