AgentPlays
Anthropic's head of international says open models carry more risk "by definition"
Guillaume Princen told Bloomberg TV at Dublin's International AI Summit that open alternatives to Anthropic's models carry more cybersecurity and misuse risk once weights are public. The comments come as open-weight options gain traction as sovereign alternatives to closed US lab systems.
How you’d use it: Weigh open-weight risk claims against closed-model safeguards.