Anthropic has given six evaluators from METR employee-like access to its models and tools, fulfilling a promise CEO Dario Amodei made last month.
Anthropic has given six evaluators from METR employee-like access to its models and tools, including company hardware and office space, so they can test unreleased models and agents.
Anthropic's draft bill would let the government block an AI model's release only if it poses a foreseeable, imminent risk of more than $1 billion in damages or killing or seriously injuring more than 50 people, according to Politico.
The bill limits intervention-worthy risks to three categories: the creation of a biological weapon, the model evading its developer's controls, or a cyberattack, Mother Jones reports.
