🌍 Impact Stories

🚔 When Research Changes Policy: The End of Predictive Policing in the Netherlands

AI accountability in action — research that reached beyond the lab.

It is not every day that academic research leaps off the page and into the heart of national policy — but here we are.

Our work through the TAILOR Network of Excellence Centres on Trustworthy AI connectivity grant, together with collaborators we, examined the Crime Anticipation System (CAS) — a predictive policing algorithm that had been in active use by the Dutch Police for over a decade.

Presented at ACM UMAP, our research surfaced critical, actionable insights about how CAS operated, the assumptions baked into its design, and the real-world consequences for communities it targeted. We were far from alone a growing body of researchers had raised alarm bells about CAS over the years, and the cumulative weight of that evidence finally tipped the scales.

In 2026, the Dutch Police announced the discontinuation of the CAS algorithm. 🎉

This is what trustworthy AI research looks like in practice: not just critique, but careful, rigorous scholarship that gives policymakers the evidence they need to act. A decade of CAS is ending — and that is a win for accountability, for fairness, and for every community that deserves policing grounded in evidence rather than algorithmic speculation.

👏 Proud of this collaboration and deeply grateful to everyone who pushed this conversation forward.

📄 Read the Paper (ACM DL)  |  📰 Read the NL Times Story

💙 A Digital Companion for Young Family Caregivers

Technology as a bridge — not a replacement — for human care and connection.

Across the Netherlands and indeed the world there is a quiet, largely invisible group carrying an extraordinary weight. Young adult caregivers, aged 18–30, make up 12–18% of all adult caregivers globally. They are students, early-career professionals, people building their lives — and simultaneously managing intense caregiving responsibilities for a family member.

They rarely make it into policy discussions. Support structures barely acknowledge they exist.

Together with Jin Huang, and building on foundational doctoral research by Dang S. at the University of Groningen, our team set out to change that. Funded by the Impact Call of the Amsterdams Universiteitsfonds (AUF), we are developing a Digital Companion for Young Adult Caregivers — an AI-powered tool designed with this community, not just for it.

The companion is being built around what these caregivers actually need: - 🧠 A specialized language module, trained on care contexts and emotional support - 🎨 A user-friendly interface, co-designed alongside young caregivers themselves - 🤝 Integration pathways with universities, healthcare providers, and caregiver organizations

The goal is not to replace human support it is to ensure that when a 22-year-old balancing exams and caregiving has a difficult moment at 11pm, there is something there for them.

"We submitted this proposal because we saw an urgent and largely unmet need. Young caregivers face significant psychological pressure and reduced quality of life, while making an indispensable contribution to our care system." — Siddharth Mehrotra

We are measuring success in terms that matter: less stress, better academic outcomes, and a healthier balance between care and self. Because caregiving should not cost someone their future.

📰 Read the Full Story (Amsterdams Universiteitsfonds)