Anthropic's revised usage policy, set to take effect on November 12, comes amid growing debate over whether advanced AI systems could possess consciousness.