Can you actually sue Anthropic over this when they clearly state that AI can make mistakes and you should double-check everything it does?