A bit about me
The short version of who I am, and why I spend time writing about this.
I research how organisations adopt and govern AI, and I build the systems too, across regulated healthcare, social services, and legal and compliance software, each under a different European regime. Mostly I care about doing it responsibly, and about the regulation around it, especially the EU AI Act.
I write here because I have had to make this work in places where getting it wrong actually costs something. I would rather share what I have learned, including the parts I got wrong, than pretend any of it is simple.
The work behind it
- Peer-reviewed research
- Papers on LLM evaluation and integration, published at RE 2026 and FLLM 2025.
- Shipped in regulated production
- AI in health, in social services, and in legal compliance, each under a different European regime.
- I refuted my own thesis
- An experiment designed to kill it did, in thirty minutes. I dropped the paper rather than publish it.
What I write about, and what I do not
Mostly: how AI actually gets adopted and governed, what responsible use really takes, where things tend to break, and what the EU AI Act means in practice rather than in theory.
Not: active clients, anything confidential, or systems that are not mine to talk about. If I have a conflict, I say so. The full version is on the Disclosures page.