OpenAI has vowed to take immediate steps to strengthen its safety protocols in a letter addressed to Canadian authorities. Anthropic said that it can't "in good conscience" comply with a Pentagon ...