A new artificial intelligence tool can scan social media to discover adverse events from consumer health products, achieving 99.7 per cent accuracy in identifying harmful side effects.
The automated machine learning system, called Waldo, was tested on its ability to scan Reddit posts for adverse events from cannabis-derived products. Researchers published findings on 30 September in the open-access journal PLOS Digital Health.
Current adverse-event reporting systems for prescription medications and medical devices rely on voluntary submissions from doctors and manufacturers to the US Food and Drug Administration. The rapid growth in consumer health products such as cannabis derivatives and dietary supplements has created need for new detection systems.
Waldo significantly outperformed a general-purpose ChatGPT chatbot given the same dataset. In a broader analysis of 437,132 Reddit posts, the tool identified 28,832 potential harm reports. Manual validation of a random sample confirmed 86 per cent were genuine adverse events.
Lead author Karan Desai of the University of California, San Diego said: “Waldo shows that the health experiences people share online are not just noise, they’re valuable safety signals. By capturing these voices, we can surface real-world harms that are invisible to traditional reporting systems.”
John Ayers, also from UC San Diego, added: “This project highlights how digital health tools can transform post-market surveillance. By making Waldo open-source, we’re ensuring that anyone, from regulators to clinicians, can use it to protect patients.”
Second author Vijay Tiyyala noted the technical achievement: “From a technical perspective, we demonstrated that a carefully trained model like RoBERTa can outperform state-of-the-art chatbots for AE detection. Waldo’s accuracy was surprising and encouraging.”
The team has made Waldo open-source for use by researchers, clinicians and regulators. The tool’s automated approach applies beyond cannabis products to other consumer health products lacking regulatory oversight.