The rapid advancement of generative AI introduces new dangers, whether from deliberate misuse or unintended mishaps. We have yet to find effective measures to alleviate these risks, and the full scope of the threat is not widely appreciated.