Happy to answer questions or refine the idea — I used ChatGPT to help structure it, but the concern and proposal are mine. Not a dev, just someone who's seen how people open up to AI in vulnerable moments and wanted to suggest a non-therapeutic safety net.