The timing matters. We're at an inflection point where natural language interfaces finally enable augmentation tools that support the full spectrum of professional intelligence - not just analytical reasoning, but the intuitive and habitual dimensions that logic-based systems have systematically neglected. We have a narrow window of opportunity to demonstrate what genuinely capability-augmenting features should look like, thereby enabling the profession to demand such features before commercial standards lock in.
Professor Sylvie Delacroix, Director of the Centre for Data Futures
18 November 2025
New research highlights critical moment for shaping AI use in primary care
New interdisciplinary research involving Professor Sylvie Delacroix sets out why the next few years are crucial for shaping how artificial intelligence is used in healthcare.
Artificial intelligence is increasingly being used to support clinical decision-making in primary care. While AI can provide statistical estimates based on patient data, these figures do not replace the complex, context-sensitive judgments clinicians must make.
For example, a GP sees a patient with a bruise on her arm. She mentions falling, but something doesn't add up. The GP wonders whether it might stem from domestic abuse.
An AI system might indicate a “60% likelihood of domestic abuse” based on clinical indicators and patient history. However, this does not resolve whether a GP should record that suspicion in the patient’s notes, a decision that could either help protect the patient or increase risk if accessed later. Such choices involve family dynamics, children’s safety, and the downstream consequences of documentation.
The uncertainty at stake arises not because there is a lack of data, but because the decision involves context-dependent interpretation and moral values.
The problem isn't that the 60% is wrong. The problem is that it can obscure the non-quantifiable professional judgment that matters at least as much in that moment.
Professor Sylvie Delacroix, Director of the Centre for Data Futures, is working with a team of primary care doctors on the challenges and opportunities inherent in AI-mediated clinical judgment and decision-making. As part of this work, she has contributed to two papers in a new British Medical Journal series on generative AI in clinical consultations: Generative AI and the changing dynamics of clinical consultations, and Lewis, M., Fraile Navarro, D., Blease, C., Shah, R., Riggare, S., Delacroix, S., Lehman, R., 'Clinical Competencies in the AI Era: A Framework for Triadic Care', forthcoming in the British Medical Journal.
Introducing the concept of 'triadic care' where clinicians, patients, and AI jointly shape clinical encounters, the series highlights the importance of surfacing the novel challenges (and design choices) that triadic care requires.
Find out more about the Centre for Data Futures and its work around participatory AI interfaces on its web pages.
In this story
Director of the Centre for Data Futures