Next Wednesday I'm in Copenhagen.
I have stage 4 colorectal cancer. I'm building FC:AI from the chemo chair. On Wednesday 20 May I'm at HIMSS Europe, on a moderated panel about how much autonomy we should hand AI in healthcare. The official title is "How Do We Define Appropriate Autonomy vs. Full Autonomy?" The room will be full of CIOs, regulators, vendors, clinicians, and a lot of people whose job titles I had to Google.
I'm the patient on the panel. That's the angle.
Who's in the room
Dr Tyson Welzel is moderating. Beside me: Sigrid Berge van Rooijen from Eir Health in the Netherlands, who runs autonomous agents in production. Dr Amr Alashkar, CIO of Cleopatra Hospitals Group in Egypt, on singleton versus multi-agent setups. HE Mubaraka Ibrahim from Emirati Health Services, the regulator on the stage. Siobhan McMahon from Dedalus has a slot earlier in the programme.
I have four minutes. After that it's a to-and-fro, and Tyson can take it anywhere he likes. He has form for it.
I'm not there to defend AI. I'm not there to be sceptical for the sake of it. I'm there to say what AI has actually done for me as a patient on chemo, and where I draw the line.
The five things I want to land
One. How I already use it. Every day. Triaging research papers I don't have the energy to read end-to-end. Summarising specialist notes so I can read oncology, genetics and surgical opinions side by side. Prepping questions before every appointment. Managing the pile of bloods, scans, pathology reports and letters that arrives faster than any one person can read it. And triaging the eight to ten people from around the world who have already reached out since I open-sourced my data.
"I cannot manage this volume of information without AI. Neither can my oncologist's team."
Two. Where I trust it, and where I don't. Anything clinical, AI proposes and a human decides. Full stop. Diagnosis, dosing, imaging, treatment changes, all of that stays supervised. Where I'm happy to give AI more rope is information work. Sorting. Pattern matching. Routing the right signal to the right specialist. The question isn't "is the model accurate enough". It's "what's the cost if it's wrong, and who is checking".
Three. The patient as connector. This is the part most people in the room won't say out loud. The future isn't AI replacing the doctor. The future is the patient becoming the connector across a fragmented system. I'm already the middleman between people far smarter than me. Oncologists, geneticists, researchers, surgeons. They mostly don't talk to each other. AI is the only reason I can hold that role without falling apart.
Four. Openness is the unlock. I've open-sourced my bloods and my scans. They sit on fcancerwith.ai/open-source-me for anyone, human or AI, to read. It's not a model. It's not a recommendation for anyone else. It's one person running one experiment in public. The replies have already reframed one treatment decision I was about to make. None of that happens if I sit quietly and wait my turn.
Five. The mic drop. Be your own advocate. Earlier than the system tells you to.
If you're going
If you're at HIMSS Europe on Wednesday, please come and find me. The panel is at 16:30 at the Bella Center. I'll be in the speakers lounge from 16:00. I would rather have one real conversation in person than another twenty LinkedIn DMs.
If you're not in Copenhagen, follow along on the FC:AI Instagram and LinkedIn through the trip. I'll post what landed, what didn't, and a recap once I'm back.
This is the first time I'll have stood on a stage and said all this out loud. I want to land it.
Be your own advocate. Earlier than the system tells you to.