The Elephant in the Room: What About AI?

Whenever I talk with advocacy groups about registries, someone asks whether AI can do this now. My answer is that AI can do a great deal of the drafting, coding, and summarizing, and I use it every day. Most of what a registry needs from its data, such as counts, descriptive tables, and trends over time, is ordinary analysis that does not require AI at all. AI works on the data a registry has collected, and it can only be as good as what is there.

A registry row is the residue of a human process. A parent found the site, read the consent, remembered a date, and typed an answer. A clinician wrote a letter or did not. A coordinator sent the reminder, or the address was out of date. An AI tool sees the residue; it cannot see what was never asked, never verified, or never recorded, and it will produce a confident answer anyway.

Each chapter in this book names what AI can do for that part of the work and what only a person can do. The point is not to avoid AI. The habits in this book are what make AI useful on registry data, and none of them can be delegated to it.