STORY / PUBLIC SOURCE RECORD
AI in clinical trials
- Published 13 February 2025; updated 25 March 2026; estimated reading time 2 minutes. The article defines AI as computer systems performing tasks associated with human intelligence, machine learning as a way for systems to learn patterns from data, and generative AI as extending this to language, code, chemistry, biology and other complex subjects.
- It defines clinical trials as studies with human participants that evaluate medical, surgical or behavioural interventions for safety and effectiveness before broad availability.
- Listed AI uses include predictive modelling to estimate possible outcomes and support decisions/risk assessment; analysing earlier studies and real-world data to optimise trial sample sizes, endpoints and methods; monitoring trial data in real time for anomalies or trends; and searching electronic health records for eligible participants.
- The page names the US National Institutes of Health’s TrialGPT as an example of a tool matching volunteers to trials, intended to improve enrolment. It says this approach can also support more diverse representation.
- Claimed benefits include faster trial operations through automation, lower costs and fewer errors, earlier detection of adverse events for quicker intervention, and stronger decisions from trial design through data interpretation, potentially supporting more targeted therapies.
- The generative-AI section points to possible trial simulation and improved predictive modelling, with ongoing research into understanding biological systems and personalising medicine. It cites the World Economic Forum’s claim that realizing generative AI’s potential could save billions and speed delivery of treatments; this is presented as a prospective estimate, not a measured outcome.
- Related links include Health and life science, Norwegian medical research, precision medicine and an AstraZeneca page on generative AI in drug discovery.