News

Trained on over 100 billion de-identified medical events, Comet aims to simulate future scenarios such as disease risks, complications, or length of stay, offering clinicians data-driven insights into what might happen next. Rather than focusing on single-disease models, the tool attempts to capture the complexity of real-world patient care, providing multiple possible pathways instead of a single prediction. Without harmonisation of data and consistent quality, there is a risk that AI-generated futures may reinforce existing blind spots or inequities rather than illuminate them. There are also practical questions: - How will clinicians interpret and act on simulated futures without overreliance on algorithmic forecasts?