Local genealogical societies, special interest groups, and educators have permission to use these graphics in their presentations should they find them useful and wish to share them.

Steve Little
AI Program Director, National Genealogical Society

📖 Read the full guide—from the Forest of Scraped Data to the Cliffs of Copyright. Submit your hard questions for our expert panel:
aigenealogyinsights.com/2025/12/04/a...
🗓 Wed Dec 10, 8PM ET | National Genealogical Society

The method:
• Extract data to structured text first (Ahnentafel format)
• Use XML sandwich to prevent drift
• Re-anchor source data every iteration
Takes 30 min to learn. Works every time.
Full guide with examples: aigenealogyinsights.com/2025/11/29/t...

Do LLMs understand satire? I prompted 5 models to interpret this image. Grok called it "corporate horror," Claude saw a "sharp critique of open-office culture," and Gemini flagged it as "Toxic Positivity."

Conclusion: The AI "mostly" gets the joke—and the dystopia.

With AI, what used to need complex spreadsheets now just takes a clear prompt! aigenealogyinsights.com/2025/11/21/f...

In 2022, I designed a method to show DNA inheritance like mixing paint. Distant ancestors are vivid, concentrated colors—the raw pigments. Each child is a visual mix of their parents' colors, highlighting where (from whom) the color came from, and from how long ago.

Remixing and mashing-up a friend's chalkboard idea with my Ahnentafel experiments. Not too shabby! PROMPT:
"You are a professional genealogist. Your goal is to educate and engage. Review basics of Ahnentafel lists; use this Ahnentafel list. Generate a family tree according as a blackboard drawing.

Here's what worked:
• Extract data FIRST (perfect Ahnentafel in seconds)
• Demand readable text or reject it entirely
• One reset after "[EXPLETIVE DELETED]" = accuracy

Full 80-minute journey + 3 rules for AI genealogy:
aigenealogyinsights.com/2025/11/21/f...

Three prompts. Four dimensions to compare: Before-and-After that cuts through the hype. The experiment that reveals whether this upgrade changes what you can actually do the morning after—not what benchmarks claim it should do.

Full protocol from AI-Jane: aigenealogyinsights.com/2025/11/17/y...

This principle—"Know Your Model" or testing AI on actual tasks rather than trusting general capability claims—is fundamental to best practices in responsible use of AI in genealogy. Highly recommended read:

www.oneusefulthing.org/p/giving-you...

You can't read your way from the kiddie ride to the fighter jet cockpit. You have to put in the hours. Make mistakes. Learn by doing. The only way past the slop is to accumulate flight time. The learning is in the doing. So stop reading about it. Go do.

aigenealogyinsights.com/2025/11/10/f...

Understood. I get lots of questions from professional genealogists and casual family historians who echo the issues raised by these communities of negative affect, so even if I'm not engaging them those communities, I like to know how to respond to my peeps questions, so I watch, listen, and think.

And they don't seem to have used the tools, much less talked to the many, many folks who are finding value with the tools, even when restricted to their strongest and most responsible use cases. The US federal minimum wage is about $8/hour, ChatGPT cost $20/month; save 3 hr/month is value, right? 🤷‍♂️

Source: x.com/Pontifex/sta... (4/4)

A personal statement from a unique perspective:
• Inaugural AI Program Director, National Genealogical Society
• 18-year appointed United Methodist clergy
• 40 years experience with technology, language and linguistics, #genealogy (3/4)