LLMs are about to enable fake digital personas, digital replicas, that the user can interact with. These will be used for self-improvement (digital coach, etc.) and for self-destruction (interactive porn, etc.). The latter is amoral, but legal, and the tech corps will exploit that mercilessly. The danger lies in our tendency to anthropomorphize LLMs simply because they quack the right way. If text-only chatbots have mesmerised people so much, imagine what chat + audio + video will do. The laws will catch up a generation later when the damage will be comparable to a forest fire.