Five Things I Learned at RootsTech 2026
The genealogy AI conversation isn't "should we?" anymore.
I’m going to be a little more direct than usual in this post.
Mark Thompson and I just delivered six sessions at RootsTech 2026 — four solo presentations and two panels on the future of AI in genealogy and ethics and standards. Somewhere in here I’m going to mention our Academy course. But these takeaways are real, they matter, and I’d be writing about them regardless.
Here’s what stood out.
1. The audience has changed — dramatically.
Mark and I poll AI usage at every session we teach. Three years ago, maybe one in eight people in a genealogy audience had heard of ChatGPT. The experience curve was skewed hard to the left — almost everyone was brand new.
That world is gone.
At RootsTech 2026, the curve looked like a proper bell curve. In two advanced classes, more than 90% of participants had written at least one line of code at some point in their lives — BASIC, Fortran, Perl, Python, whatever it was, even the classic first-day “Hello, World” exercise. These aren’t people starting from scratch. In other words, even a smidgen of technical knowledge, even if it’s forty years old, can be powerfully leveraged with AI. Today.
At the bleeding edge, some members of our community are now approaching 3,000 hours of hands-on AI experience.
And at the same time, there are many, many people with NO experience now ready to start learning. (And their timing is perfect—it’s never been easier to get started, even if you’re starting from the very beginning.)
The conversation isn’t “what is AI?” anymore. It’s “how do I use it well?”
2. The vendors noticed.
This wasn’t a year of cautious experiments. The major platforms showed up with production-ready AI features.
MyHeritage launched Scribe AI just before the conference — a tool that transcribes, translates, and interprets historical documents, and was being demonstrated live on the expo floor.[1] Ancestry showcased AI Stories, which turns a census record or draft card into a narrated audio story.[2] And FamilySearch moved its full-text search — now spanning over two billion document images — out of Labs and into production, alongside a new AI Research Assistant.[3]
When FamilySearch, Ancestry, and MyHeritage are all shipping AI features in the same quarter, I don’t think anyone at RootsTech was still asking whether AI belongs in genealogy. The question is whether you understand what these tools are doing well enough to trust their output — or catch their mistakes.
3. The privacy conversation got real.
The Church News quoted me on this one: the responsible use of AI in genealogy isn’t optional. It’s foundational.[4]
At the Thursday panel on “Responsible Use of AI,” the room was full and the questions were sharp. People aren’t afraid of AI anymore — they’re afraid of using it carelessly. They want guardrails. They want frameworks. The Coalition for Responsible AI for Genealogy, host of this panel, has done this preliminary work.
We offered ours: three layers that build on each other.
Layer 1: The Human Rule — AI advises, humans decide. Every claim gets verified against original sources. This is the on-ramp.[5]
Layer 2: Words, Not Facts — Language models produce language, not verified facts. Understanding this single distinction prevents most AI-generated genealogy errors.
Layer 3: ADPEC — Accuracy, Disclosure, Privacy, Education, Compliance. The full ethical framework for responsible AI use in genealogical research.[6]
These aren’t abstract principles. They’re practical checkpoints that work whether you’re using ChatGPT, Gemini, Claude, or the AI features now built into the platforms you already use.
4. Your money’s better spent learning to make a tool than to buy a tool.
The major platforms weren’t the only ones at RootsTech with AI features. The expo hall was also full of smaller apps — $5, $10, $20 a month — each doing one narrow thing. Transcribe a document. Enhance a photo. Generate a biography. Alton Brown used to call kitchen gadgets like this “uni-taskers”: expensive single-purpose devices when a good knife and a hot pan would do.
The AI version of a good knife and a hot pan is learning to use a general-purpose tool — ChatGPT, Claude, Gemini — and giving it clear instructions in the context of your actual research. That’s what my Friday session on “User-Created AI Tools for Family History” was about, and it drew some of the most engaged questions of the conference. People are building their own workflows, their own prompts, their own lightweight applications. They’re making Swiss Army knives instead of filling a drawer with uni-taskers.
Is buying a tool sometimes the right call? Of course. The real question is: do you have more time or more money? But you can’t make that decision well unless you know what you’re capable of building yourself.
5. Nobody needs to pay for AI education.
There is no piece of AI knowledge that you can’t find for free somewhere. YouTube tutorials, blog posts, forum discussions — it’s all out there.
So why does our Academy exist?
Because information isn’t the bottleneck. Context is. What Mark and I offer isn’t knowledge you can’t find elsewhere. It’s four things that are genuinely hard to assemble on your own:
A genealogical context for learning AI — not generic productivity tips, but AI skills grounded in the actual tasks genealogists face every day
A thoughtful, proven curriculum — structured learning that builds week over week, not a random collection of tricks
Current tools and topics — nothing gets even remotely stale; we update materials between every cohort
And most especially: a continuing community — the people in your cohort become colleagues. The learning doesn’t stop when the course ends.
That’s the value proposition. Not exclusive knowledge. Not another uni-tasker. A path through the noise, with people who understand why it matters.
If you haven’t written your first line of code yet — that’s exactly who Level I is for.
Here’s the plug.
Our Level I course — Introduction to Family History AI — starts March 17. Five weeks. Ten sessions. Everything from prompting to image analysis to AI-enhanced research workflows.
To celebrate RootsTech, we’re offering 15% off with code ROOTSTECH15OFF — $212 instead of $249. The discount expires when class starts.
Register here: https://tixoom.app/fhaishow/1ycwuhd5
No prerequisites beyond an interest in genealogy and a desire to understand what AI can — and can’t — do for your research.
And if you’re one of those 3,000-hour users? Level III starts May 5. That one’s for you.
Notes
[1] “Introducing Scribe AI,” MyHeritage Blog, 4 March 2026. https://blog.myheritage.com/2026/03/introducing-scribe-ai/ — See also the BusinessWire press release: https://www.businesswire.com/news/home/20260304679582/en/ (accessed 10 March 2026).
[2] “Ancestry Brings Family History to Life with New AI-Powered Stories,” Ancestry Corporate Blog, 12 December 2025. https://www.ancestry.com/corporate/blog/ancestry-brings-family-history-to-life-with-new-ai-powered-stori — AI Stories launched in beta December 2025 and was showcased at RootsTech 2026 during Crista Cowan’s “What’s New at Ancestry” session (accessed 10 March 2026). See also Sarah Needleman, “Ancestry’s new AI feature narrates ancestors’ stories,” Semafor, 12 December 2025. https://www.semafor.com/article/12/12/2025/ancestrys-new-ai-feature-narrates-ancestors-stories
[3] “AI and Genealogy: Advancements You Can Use,” FamilySearch Blog, 2026. https://www.familysearch.org/en/blog/ai-developments-genealogy — FamilySearch’s own published figure is “nearly 2 billion” records as of January 2026 (https://www.familysearch.org/en/blog/what-is-full-text-search). The higher figure was reported in conference coverage by Randy Seaver, “Randy (Not) at RootsTech 2026 — Day 1,” Genea-Musings, 5 March 2026. https://www.geneamusings.com/2026/03/randy-not-at-rootstech-2026-day-1.html (accessed 10 March 2026).
[4] “RootsTech 2026 researchers share 5 principles to use AI responsibly,” Church News, 6 March 2026. https://www.thechurchnews.com/living-faith/2026/03/06/researchers-provide-guidelines-for-responsible-ai-usage-family-history-rootstech/ (accessed 10 March 2026).
[5] “Avoiding Extremism,” Steve Little, , AI Genealogy Insights, 9 June 2024. https://aigenealogyinsights.com/2024/06/09/avoiding-extremism-the-use-and-disclosure-of-ai-in-genealogy/ (accessed 10 March 2026).
[6] “Guiding Principles for Responsible AI in Genealogy,” Coalition for Responsible AI in Genealogy, 2025. https://craigen.org/ (accessed 10 March 2026).