by Dr Shereen Fouad (Senior Lecturer in Applied AI and Robotics) and Professor Aniko Ekart (Co-Lead at AI.RDN+ and Director of the Aston Centre for Artificial Intelligence Research and Application)
9th June 2026
Designing a trustworthy AI education framework in healthcare has never been just a curriculum design exercise; it has been an exercise in coalition‑building. Our recent preprint on a tiered AI capability framework for medical imaging, available here, sits at that intersection between service pressures, regulatory complexity, and the ambitions of a research ecosystem that is racing ahead.
We started from a simple observation: AI is already present in imaging workflows, yet many professionals feel under‑prepared to question, implement, or govern these tools. Rather than designing a course in isolation, we convened a co‑design workshop, available here, that brought together radiologists, radiographers, physicists, data scientists, managers, and patients. That mix was crucial. Clinicians surfaced anxieties about explainability and workflow disruption; developers stressed data quality and validation; patients asked about trust, consent, and “who is accountable when things go wrong.” These perspectives became guardrails for the framework, ensuring it spoke not only to technical competence but to lived realities on the ground.
The result was a tiered architecture: foundational AI literacy for the wider workforce, an imaging‑focused proficiency pathway for those working directly with AI tools, and a policy and governance pathway for leaders. This structure is deliberately asymmetrical. Not everyone needs to tune models, but everyone needs a shared language for concepts such as bias, generalisability, and human oversight. In practice, that meant moving away from “AI for radiologists” towards “AI with radiologists, nurses, managers, and patients in the same conceptual frame.” It also meant centring responsible innovation – safety, accountability, and regulation -rather than viewing ethics as a bolt‑on at the end of a technical syllabus.
Shereen Fouad the project principal investigator, commented on this outcome, “Building on this project, we now plan to embed the framework as a standing element of Aston’s digital health research and training offer (here), including dedicated professional development for new NIHR Academic Clinical Fellows joining Aston Medical School’s expanding programme. These colleagues sit at the interface of research and frontline care, and equipping them with robust, governance‑aware AI capability will be central to how we grow a responsible, clinically led AI research ecosystem in the years ahead.”
Aniko Ekart, the project co-investigator, further commented on this outcome, “What has stayed with me most is the impact these courses are already having on the medical workforce. Seeing clinicians arrive wary of “black box” systems and leave able to question, challenge, and help shape AI projects in their own services has been incredibly powerful. Several participants have gone on to lead local AI governance groups, redesign imaging workflows, or push back on unsafe implementations – exactly the kind of critical, constructive engagement we hoped to enable. I am looking forward to adapting the learning from this project to AI.RDN+, aiming to achieve similar impact within the doctoral ecosystem.”
Looking ahead, the framework offers more than a set of AI professional training courses. It is becoming a shared map that healthcare organisations, educators, and doctoral programmes can adapt together. If we want AI research to translate safely and meaningfully into medical practice, we need professionals and researchers who are fluent not only in models, but in each other’s worlds. This work is one attempt to build that common ground.
****************************************
If you would like to share your thoughts on the article or share the article itself on LinkedIn, please click here.