\n<\/p>\n
Using artificial intelligence chatbots to seek medical advice can be dangerous because of their tendency to provide inaccurate and inconsistent information, according to a new study by researchers at the University of Oxford.<\/p>\n
The research, led by scientists from the Oxford Internet Institute and the Nuffield Department of Primary Care Health Sciences, was published in the journal Nature Medicine<\/em>.<\/p>\n Researchers asked nearly 1,300 participants to identify possible health conditions and recommend next steps across a range of medical scenarios. Some participants used large language model software to generate potential diagnoses and advice, while others relied on traditional methods, including consulting a general practitioner.<\/p>\n After evaluating the responses, the researchers found that AI tools often delivered a mix of accurate and inaccurate information that users struggled to differentiate. The study concluded that while chatbots can perform well on standardized medical knowledge tests, their use as a clinical support tool poses risks to individuals seeking help for personal symptoms.<\/p>\n Dr. Rebecca Payne, a co-author of the study and a general practitioner, said the findings show that AI is not yet ready to replace physicians.<\/p>\n \u201cPatients need to be aware that asking a large language model about their symptoms can be dangerous,\u201d Payne said. \u201cThese systems can give incorrect diagnoses and fail to recognize when urgent medical attention is required.\u201d<\/p>\n The researchers noted that the challenge lies not only in medical knowledge but also in human interaction. Lead author Andrew Bean of the Oxford Internet Institute said that even top-performing large language models struggle in real-world interactions involving nuanced, high-stakes decisions.<\/p>\n \u201cThese findings highlight the difficulty of building AI systems that can genuinely support people in sensitive areas like health,\u201d Payne said.<\/p>\n Bean said the research aims to contribute to the development of safer and more reliable AI systems, particularly as interest in using chatbots for health advice continues to grow.<\/p>\n<\/p><\/div>\n \n