00:00It's Benzinga, bringing Wall Street to Main Street.
00:02Anthropic announced Thursday that it is partnering with nonprofits METR and SecureBio
00:08to embed independent evaluators within the company to assess artificial intelligence safety practices.
00:13The company is developing a three-tier system that grants evaluators varying levels of access and publication rights.
00:19Anthropic has granted six METR employees access to internal resources to assess its risks,
00:25incidents, and practices and publish their findings without funding the organization's work.
00:30SecureBio has more limited access and is reviewing chemical and biological weapons risks.
00:35Anthropic is also discussing potential partnerships with Hugging Face and expanding its work with Accenture.
00:40The initiative follows CEO Dario Amode's proposal last month to slow AI model development
00:45and independently verify safety practices.
00:48Anthropic said it believes all frontier AI companies should adopt independent evaluations.
00:53For all things money, visit Benzinga.com.
