Anthropic absent from open AI alliance letter, Amodei denies ban push

Key Points:

• Anthropic CEO Dario Amodei clarified that his company does not support a ban on open-weight AI models, countering claims linked to Anthropic's absence from a recent letter signed by leading AI firms advocating for open AI collaboration.
• The letter, supported by companies like Nvidia and Palantir, promotes the Open Secure AI Alliance to develop and share open technologies aimed at improving AI security, acknowledging risks from both open and closed AI models.
• Amodei emphasized that while he values open-weight models as a public good for expanding AI access and competition, he disagrees with the notion that open models inherently make it easier to develop safeguards or benefit defenders more than attackers.
• He advocates focusing on preventing powerful AI chips from authoritarian misuse, stopping large-scale unauthorized model replication, and enforcing safety testing for all capable AI models, rather than banning open-weight models outright.
• Palantir CEO Alex Karp, a signatory of the open-weight support letter, warned against restrictive AI regulations that could stifle competition and innovation, stressing the importance of competing openly to maintain leadership in AI technology.