Examine strategies for validating, monitoring, and safely deploying configurable AI agents to ensure compliance and performance in clinical trials.
In a recent video interview with Applied Clinical Trials, Michelle Longmire, MD, co-founder and CEO of Medable, discussed the role of agentic artificial intelligence (AI) in clinical trials. She highlighted that 45% of clinical development time is wasted due to delays and siloed data. Agentic AI aims to reduce this by leveraging technology in areas previously untouched, freeing up human resources for strategic work. The human-in-the-loop concept ensures that AI supports, rather than replaces, human decision-making. Integration of AI across legacy systems is facilitated by model context protocols, reducing manual effort. Automation handles routine tasks, while critical decisions remain human-controlled. No-code platforms for AI tools require clear job descriptions and validation similar to human training.
ACT: No-code platforms for building AI tools are becoming more common. From a clinical operations perspective, what safeguards or governance structures should be in place when teams configure their own agents for trial use?
Longmire: The configuration of agents is really important, and I'd say the first way to think about it is like you're writing a job description for a human. The more we begin to think about agentic AI as a true human collaborator, I think we're operating in a zone where we're getting the leverage that the system technology can provide. When we instruct, mentor, teach, train humans, we want to not only provide really clear direction and what that needs to be, we also want to validate that the human can do those things effectively, and agentic AI is really no different. We want to start with a job description. We want to start with the goal orientation, the job to be done, and then we want to ensure that the agent actually operates effectively and performs well against those critical jobs. The more we think about it like humans, the bigger bite of the apple we're able to take. Just like a human, I'm a physician. The first couple times I'm in the operating room, I'm not touching anything. I'm just observing. Now, as I go through my residency, I become board certified, I'm taking care of many more patients, there are many layers of validation that happen to ensure that I am appropriately trained and credentialed to do that work.
