Anthropic has raised significant concerns about the risks associated with AI systems, including their potential autonomy, legal implications, and the impact on employee safety decisions, as it prepares for its IPO.