Twenty-five US technology companies, including Nvidia, Microsoft, IBM, and Meta, signed an open letter this week calling for open weighted AI models. The open letter stated that although open weights have risks such as difficulties in controlling derivative versions, the closed model also has security risks. An open ecosystem can mobilize a large number of researchers to investigate vulnerabilities and improve security protection, and is an important path for achieving artificial intelligence security. Currently, 50 companies/organizations have signed the letter. Newly added companies include Google, AMD, Cloudflare, Block, and Ollama, as well as OpenAI, Y Combinator, GitHub, and Cisco as previously reported. Notably, Amazon and Anthropic have yet to sign this statement, which has attracted attention from the outside world. “Forbes” believes that the reason for this is commercial interests, and Amazon is Anthropic's biggest investor. Furthermore, Anthropic has only provided closed source AI model services from the beginning to the end, and the company has always maintained an “AI security” stance.

Twenty-five US technology companies, including Nvidia, Microsoft, IBM, and Meta, signed an open letter this week calling for open weighted AI models. The open letter stated that although open weights have risks such as difficulties in controlling derivative versions, the closed model also has security risks. An open ecosystem can mobilize a large number of researchers to investigate vulnerabilities and improve security protection, and is an important path for achieving artificial intelligence security. Currently, 50 companies/organizations have signed the letter. Newly added companies include Google, AMD, Cloudflare, Block, and Ollama, as well as OpenAI, Y Combinator, GitHub, and Cisco as previously reported. Notably, Amazon and Anthropic have yet to sign this statement, which has attracted attention from the outside world. “Forbes” believes that the reason for this is commercial interests, and Amazon is Anthropic's biggest investor. Furthermore, Anthropic has only provided closed source AI model services from the beginning to the end, and the company has always maintained an “AI security” stance.