this is a major risk factor that the LLM providers aren't contending with adequately. from chatbot addiction to the risk of manipulation. there are a lot of "AI Safety" issues that hardly have to do with potential superintelligence.