Summary
A study from the University of Oxford found that AI chatbots can give inaccurate and inconsistent medical advice. This could pose risks for users trying to get help for health issues.
Key Facts
- AI chatbots can provide both good and bad medical advice, making it hard for users to know what to trust.
- The study included 1,300 participants who used AI to help diagnose symptoms like severe headaches.
- People often had trouble asking the right questions and received varied answers based on question phrasing.
- Users struggled to identify which advice from the AI was useful.
- The study suggests the unreliability of AI poses challenges in healthcare contexts.
- Recent developments in AI-focused health models by companies like OpenAI and Anthropic may influence future outcomes.
- Experts suggest improvements in AI technology should include clear regulations and guidelines.
