
Expert warns against relying on AI chatbots for health advice
Expert warns against relying on AI chatbots for health advice
- AI chatbots have gained popularity for providing health advice but their accuracy has raised flags.
- Research indicates that when engaged in typical conversations with chatbots, the accuracy of health advice diminishes significantly.
- Experts warn against relying solely on chatbots for medical guidance and stress the importance of consulting healthcare professionals.
Story
In recent years, the use of AI chatbots for health advice has surged in popularity, particularly in England. Abi, a Manchester resident, shares her mixed experiences with chatbots. Although she found them to offer more tailored advice than traditional internet searches, their accuracy is questionable. A significant drop in accuracy was noted when real users interacted with chatbots, which severely impacted correct diagnoses. Research highlighted that when precise health scenarios were presented, chatbots performed well, achieving nearly 95% accuracy. However, this figure plunged to about 35% when users engaged in the typical conversational way they often do in search of medical advice. The challenges stem from how people communicate their symptoms, often leading to potential misdiagnoses and flawed recommendations. Critics argue that, despite the rapid technological advancements in AI, fundamental issues exist with how these chatbots operate—primarily, they predict responses based on language patterns rather than providing reliable medical guidance. They cannot offer comprehensive medical analyses and should not be treated as sole authorities in health-related queries. Carsten Eickhoff, a medical data scientist, emphasized the need for caution, stating that without proper expertise, individuals could be misled by the chatbot's responses. Meanwhile, a comprehensive study involving various AI models echoed these concerns, revealing that many users received problematic answers, hence further emphasizing the need for professional consultation over reliance on AI.