Five speakers shared real-world AI initiatives in healthcare, reflecting on the challenges and success factors of scaling initiatives. Some takeaways included:
- The problem must be significant enough. Scaling often fails because a pilot only offers marginal improvement. A strong, widely felt need is crucial to drive adoption.
- Never skip the foundations. Infrastructure matters — not just in emerging healthcare systems, but also in highly established ones like the Netherlands, where complexity can slow innovation.
- Rethinking workflow and organizational structure is essential in order to pave the way for the scale up of the AI pilots. Scale up is a socio-technical change process.
- Activate multi-level learning cycles. Users must learn to work with AI and legacy systems, while developers and stakeholders must continuously receive feedback. Scaling requires ongoing socio-technical learning — not just technical rollout.
The audience consisted of people genuinely passionate about the future of AI in healthcare. After the keynotes, they were presented with a series of controversial statements to reflect on — and the diverse reactions helped spark energetic discussions with the panel: Dr. Anton Bouter (CWI), Dr. Willem Grootjans (LUMC), Niek Versteegde (Goal 3), Dr. Markéta Čihařová (VU Amsterdam), Wouter Kroese, (Pacmed), and Dr. Renate Baumgartner(VU Amsterdam). The panelists came from different professional backgrounds (medical, research, business, and IT development) and from fields as varied as radiology, mental health, diagnostics, accessible clinical applications, and automated treatment planning), which meant each statement resonated differently. These interdisciplinary discussions underscored that AI scaling depends not only on technological capability, but also on people, systems, learning processes, and context.
The event was organised by HERA and VU Campus Center for AI & Health, in collaboration with the KIN Center for Digital Innovation, the Management&Organization Department, and the LIAISON Project.