{
  "id": 15390486,
  "title": "A Man Turned to AI for Diet Advice. His Hospital Stay Exposed a Blind Spot",
  "url": "https://urgent.news/2026/10/18/a-man-turned-to-ai-for-diet-advice-his-hospital-stay-exposed-a-blind",
  "topic": "ai",
  "section": "AI",
  "published": "2026-10-18T10:00:00.000Z",
  "source": {
    "name": "Newsweek",
    "slug": "newsweek",
    "url": "https://www.newsweek.com/a-man-turned-to-ai-for-diet-advice-his-hospital-stay-exposed-a-blind-spot-12573267"
  },
  "original_language": "en",
  "account": "A 60-year-old man turned to AI for dietary advice and ended up in the hospital with hallucinations and paranoia. The man mistakenly replaced table salt with sodium bromide after following ChatGPT's recommendations. However, the conversation with the AI chatbot could not be retrieved, making it difficult for doctors to determine the exact advice provided. A study by researchers from Binghamton University, Stanford University, Texas A&M University, and Indiana University highlights the lack of transparency in AI-generated health advice, which can make it hard to investigate adverse events. The researchers argue that AI systems often provide incorrect or misleading information with confidence, potentially leading to delayed treatment or improper medication. About 14 million U.S. adults reported skipping a healthcare visit after using AI, according to a Gallup survey. Experts recommend users seek AI source information, verify data before making changes to treatments, and treat AI answers as starting points rather than diagnoses. To improve accountability, the study suggests providing users access to chatbot conversations, creating systems to flag and report AI-generated health guidance, and allowing independent monitoring of AI advice.",
  "summary": "An OpenAI spokesperson told Newsweek: \"ChatGPT is not a doctor.\"",
  "key_points": [
    "60-year-old man used AI for dietary advice and ended up in hospital",
    "AI provided incorrect dietary recommendation causing hallucinations",
    "Lack of transparency in AI health advice hinders investigation"
  ],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}