Anthropic and OpenAI answer questions on evaluator publication rights

Reuters reported on October 13 that Anthropic and OpenAI sought significant control over what outside safety evaluators can publish, with two people saying both labs still wanted strict editorial control as of October 5. Anthropic told Reuters that existing evaluators control their methods, criteria, conclusions, and publication, while it reserves a right to comment on and approve reports from new evaluator firms. OpenAI said it does not seek approval over full reports and that its practice is to redact confidential information, which evaluators may flag.