Imagine your doctor uses an AI tool to help work out what is wrong with you. It gives a confident answer, your doctor trusts it, and the answer is wrong. You are the one who gets hurt, and when the claim is filed the company that built the AI is nowhere in the paperwork. Your doctor is.
Here's why that's a big deal: when a defective car or a bad drug hurts someone, the maker pays. The law has never decided that AI counts as a product, so the companies behind it stay largely out of reach, and the harm slides downhill to the nearest human signature. The Medical Protection Society calls the clinician the liability sink, and says more than 350,000 of them are now exposed. A doctor cannot open the model and inspect its reasoning. They can only trust the output or ignore it, and they carry the legal risk either way.
So what does that mean for you? The next time a doctor reaches for AI to help treat you, remember that if it goes wrong, the person across the desk is the one holding the bag. Until liability is assigned when these tools are deployed, instead of discovered in court after the harm, every confident answer comes with a quiet question about who pays for a mistake.