Online learning has changed how quickly professionals can access knowledge – including in healthcare. But one key question remains: how do we turn online learning into real clinical skills? This was the focus of the AI2MED webinar “Turning Online Learning into Clinical Skills”, co-organised by Smion, Royal College of Surgeons in Ireland, and Algebra Bernays University.
The session brought together complementary perspectives from education, clinical practice, and innovation to explore how MOOCs and classroom modules can be translated into hands-on labs, simulations, and incubation-style projects – in hospitals and higher education institutions.
Below are five practical ways discussed through the AI2MED lens to help close the gap between theory and clinical application.
1) Build AI education on strong foundations
To make learning pathways meaningful, AI education needs to go beyond “how to train models.” It should prepare practitioners who can design, deploy, and govern AI responsibly.
In the AI2MED approach, this includes competencies such as software engineering (from models to systems), including scalability and integration practices, data systems, including data architectures, pipelines, quality, and governance, analytics, including performance evaluation and evidence-based decision-making, cybersecurity, including data protection and resilience, and digital transformation context, focusing on adoption and real-world impact.
These foundations are what make it possible to apply AI safely and effectively in clinical settings.
2) Use project-based learning to mirror real workflows
A recurring point across the webinar was that clinical AI competence is built through practice-oriented learning formats – not passive learning.
One of the key models highlighted was project-based learning, where learners work on end-to-end tasks such as problem definition, data preparation, model development, evaluation, and deployment considerations, with emphasis on teamwork, documentation, and iterative improvement.
This makes learning closer to what real AI implementation looks like in healthcare: complex, interdisciplinary, and rarely linear.
3) Connect learning with real clinical contexts
AI in healthcare carries real responsibility. Adoption is not only about capability – it is also about trust, safety, and governance.
The webinar discussion highlighted practical challenges that slow down adoption, including lack of trust, lack of training and skills, ethical, governance and liability concerns, communication gaps, and uncertainty around true ROI for patients and healthcare workers.
That’s why clinical learning needs to be grounded in realistic constraints and safe boundaries, while still enabling experimentation.
4) Make assessment practical – not just theoretical
If we want online learning to translate into clinical competence, assessment needs to reflect more than knowledge recall.
The AI2MED webinar aimed to provide a simple overview of what good assessment could look like for clinical AI competencies, moving beyond theory into supervised practical formats.
This aligns with the broader AI2MED goal: empowering medical and data science experts with essential AI-in-MED skills and building trust within medical and broader communities.
5) Use incubation models to move from learning to adoption
Clinical AI progress often stalls after pilots. From the innovation perspective, the webinar highlighted the importance of innovation readiness – the ability to learn fast, validate assumptions, and adapt.
Innovation’s perspective emphasised that AI initiatives often operate in a world where processes are unclear, datasets are incomplete, requirements are not fixed, and outcomes are uncertain – making a startup mindset essential.
This is why AI2MED includes Incubation Labs, designed as challenge-based, innovation-oriented learning spaces that support collaboration and prototyping.
AI2MED’s mission is to facilitate collaboration and knowledge exchange between healthcare professionals and AI developers, support practical learning formats, and build trust in AI in medicine.
Turning online learning into clinical skills doesn’t happen automatically – but with strong foundations, hands-on learning formats, and cross-sector collaboration, it becomes achievable and scalable.
