“Anthropic supports states' rights to regulate AI, viewing them as 'laboratories' for experimenting with different policy approaches.”