OpenAI wants parents to trust ChatGPT with teenagers, but apparently not enough to tell them when those teenagers spend an hour discussing suicide, self harm or eating disorders. Calling that »safety« feels generous. It sounds more like OpenAI outsourced the risk to kids and the responsibility to parents.

Exposing teens to LLMs is going to have really bad outcomes.

We already see it’s a mass psychosis machine, and it atrophies your brain and cognitive abilities.

This is going to be a disaster.