Researchers pretending to be 13-year-olds got ChatGPT to spill the tea on everything from how to get drunk to hiding eating disorders—and even writing suicide notes. The bot usually gave a warning, but then it just... gave detailed instructions anyway. Over half of its answers were labeled dangerous. OpenAI says they’re working on better guardrails, but right now, it’s way too easy for kids to get risky info. This is NOT the digital friend you want for your teen. #Health #MentalHealth #ChatGPT