How AI and Digital Twins Are Transforming Clinical Trials

This is the Applied Clinical Trials Brief—your fast track to the latest insights in clinical research operations. In under three minutes, we’ll recap top stories, highlight expert interviews, and keep you current on what’s moving the industry. Let’s get into it.

On this episode, we’re sharing highlights from a recent conversation with Jon Walsh, founder and chief scientific officer of Unlearn.AI, on how artificial intelligence and digital twins are transforming clinical research.

Walsh explained that while AI-designed therapies enter trials through the same clinical testing framework as traditional drugs, they often require closer attention to biomarker signals and computational safety models early on. Digital twins—probabilistic models that simulate how individual patients would respond to standard of care—help sponsors detect efficacy signals faster and make more informed go/no-go decisions.

He described how digital twins work: rather than adding new patients, they provide additional insight into those already enrolled. In randomized studies, they improve statistical power and confidence in results, while in early-phase or open-label trials, they can serve as virtual comparator arms when control groups aren’t feasible.

Regulators are already shaping the path forward. Walsh pointed to recent FDA and EMA guidance, which emphasize defining a model’s context of use, evaluating risks, and ensuring transparency. Digital twins designed for randomized trials have already aligned with this framework, and agencies have provided positive feedback on their use.

Transparency and interpretability are critical. Walsh stressed that models must be built and applied in a traceable way, with clear data custody to avoid unblinding. Tools that explain how predictions are made and which factors drive outcomes ensure the evidence is both trustworthy and clinically meaningful.

Looking ahead, Walsh expects broader adoption of digital twins across therapeutic areas. In randomized studies, they could reduce the size of control arms and make trials more patient-centric, while in adaptive and open-label designs, they could accelerate timelines and expand patient access to new treatments—all while maintaining the quality of evidence regulators require.

For more on this and other developments in clinical research, visit us at appliedclinicaltrialsonline.com. Thanks for listening to the Applied Clinical Trials Brief.