Study reveals AI chatbots struggle to give people useful health advice

A recent study led by the Oxford Internet Institute has revealed that artificial intelligence (AI) chatbots struggle to provide people with useful health advice.

Despite the growing trend of using AI-powered chatbots such as OpenAI’s ChatGPT for medical self-diagnosis, the study has found that participants who used chatbots were less likely to identify relevant health conditions and more likely to underestimate their severity.

It should be noted that the study involved around 1,300 participants in the UK who were given medical scenarios and tasked with identifying potential health conditions using chatbots and their methods.

The results showed that the AI chatbots' responses often combined good and poor recommendations, and participants frequently omitted key details or received answers that were difficult to interpret.

The study highlighted the complexity of interacting with human users and the need for better evaluation methods.

The study's findings also have implications for the growing use of AI in healthcare.

While tech companies are increasingly pushing AI to improve health outcomes, professionals and patients are mixed about whether AI is ready for higher-risk health applications.

Experts have recommended that people should rely on trusted sources of information for healthcare decisions.

Moreover, chatbot systems should be tested in the real world before being deployed, similar to clinical trials for new medications.