New research indicates that AI chatbots are no more effective than search engines for medical self-diagnosis and may pose risks due to communication gaps and incorrect advice.
A study published in Nature Medicine reveals that AI chatbots do not provide superior health advice compared to traditional internet searches. Despite high scores on medical exams, chatbots failed to help users accurately identify ailments or determine necessary care. Research showed patients correctly identified health problems only a third of the time when using AI tools.
“Despite all the hype, AI just isn’t ready to take on the role of the physician,” stated co-author Rebecca Payne from Oxford University. The study involved 1,300 participants using models like GPT-4o and Llama 3. Findings suggest a communication gap exists, as real-world users often fail to provide chatbots with all the relevant medical information needed for accuracy.
Experts warn that relying on these models can lead to dangerous outcomes. “Patients need to be aware that asking a large language model about their symptoms can be dangerous, giving wrong diagnoses and failing to recognise when urgent help is needed,” Payne added. With one in six U.S. adults using AI for health queries, the risk of misunderstanding remains high.
NEWS NOW:
- Nicki Minaj responds to Newsom’s ‘I’m a 960 SAT guy. I can’t read’
- WATCH: Fareed Zakaria’s take on blue cities
- Kash Patel responds following backlash for celebrating with US hockey team while on work trip
- Cartels attack civilian targets across Mexico after U.S.-Mexico mission kills drug leader ‘El Mencho’