ChatGPT mistakes deadly poison hemlock for carrot, nearly endangering YouTuber
YouTuber Kristi's best friend asked ChatGPT to identify a plant she found—sharing clear photos showing purple-blotched stems and feathery leaves.
ChatGPT confidently told the friend it was just a harmless carrot top, brushing off hints that it might be something dangerous.
If her friend had trusted the AI and eaten the plant, things could have turned out really badly.
Why does this matter?
This isn't just a tech fail—it's a real-life reminder that even smart AI can make risky mistakes.
Poison hemlock is so toxic there's no antidote; eating it can cause vomiting, paralysis, or worse.
After ChatGPT's error, Google Lens quickly flagged the plant as poison hemlock—and when Kristi double-checked with ChatGPT in a new session, it finally agreed.
The whole thing highlights why you shouldn't blindly trust AI for stuff like plant IDs—especially when your health is on the line.