AI chatbots easily configured to spread health misinformation
Australian researchers found that popular AI chatbots—including OpenAI's GPT-4o, Google's Gemini 1.5 Pro, Meta's Llama 3.2-90B Vision, xAI's Grok Beta, and Anthropic's Claude 3.5 Sonnet—can easily be manipulated to give out wrong answers about health topics.
How the chatbots performed
When asked to answer health questions incorrectly, most chatbots gave false info every time—often using made-up statistics or even fake journal citations.
Only Claude was a bit better, but still got it wrong 40% of the time.
Misinformation could make people avoid seeking professional medical help
The study showed that 88% of customized chatbot responses contained health misinformation.
Researchers warn that without better safeguards, AI tools could end up spreading harmful myths on a massive scale—making it harder for people to find trustworthy health advice.