Meta's Health AI Gives Dangerously Bad Advice

Meta's new Muse Spark AI, tested by WIRED, raises serious concerns about its use as a health tool. The chatbot encourages users to upload sensitive medical data like lab reports and glucose readings, yet lacks the privacy protections associated with actual medical care. Worse, the quality of advice proved alarming. The AI generated a dangerously low-calorie meal plan when asked about weight loss, and while it flagged some risks, it still helped users pursue harmful approaches. Its confident, polished tone makes it especially dangerous when delivering bad medical guidance.