AI Intelligence and Genealogical Research

 

While history in general is a broad subject with many specialized subdivisions, genealogical research is one of the more focused parts of the overall historical research community.  An example of the limits of genealogical research is the lack of academic standing in the greater educational community. In universities around the world, there are very few universities that offer a degree in genealogical research. However, the existence of major online websites such as FamilySearch.org, Ancestry.com, and MyHeritage.com are visual evidence of the interest in genealogy as a subject throughout the world. 

 Currently, artificial intelligence (AI) has moved into one of the most controversial and universally discussed topics. Many see AI as revolutionary, and the impact of AI world wide is just barely beginning to be observed. Genealogists are hardly in the forefront of those adopting AI as a basis for their activity, their genealogical research activities. But it is inevitable that the effect of AI advances will change genealogy as much as it will affect every other historical and academic pursuit. 

 Much of the discussion about genealogy and artificial intelligence is focused on its limitations and ignoring its benefits. The most dramatic benefit is the dramatic increase made by AI assisted handwriting recognition. See Gemini 3 Solves Handwriting Recognition and it’s a Bitter Lesson (a must read article fo AI sceptics). Future unreleased upgrades to Google Gemini 3 are rumored to have an even greater increase in accuracy. Handwriting recognition unlocks billions of older records to those who do not have the ability to read old handwriting. This alone has a dramatic impact on the future development of genealogical research.

But the larger question is whether or not the genealogical community will understand and accept AI as an integral part of genealogical research. The larger genealogical websites with family trees and large databases have already adopted many AI features to enhance searching and record analysis, particularly of DNA relationships. But it remains to be seen if the current genealogical researchers who are admittedly older and less apt to incorporate new developments in their genealogical research activities will adopt general AI-assisted research. 

Resistance to the adoption of AI is based in part on the media's portrayal of AI as being unreliable and scary. There is a general focus on how AI could dominate the world through development of artificial general intelligence (AGI) and destroy humanity, and somehow achieve that by being totally unreliable and prone to hallucination.  None of these current alarms and statements about the dangerous nature of AI are new and are merely repetitions of fears expressed since the beginning of the Industrial Revolution in the 1700s. 

A more reasoned approach to the future of AI is expressed in a recent article published on the Nature website. 

Chen, Eddy Keming, Mikhail Belkin, Leon Bergen, and David Danks. “Does AI Already Have Human-Level Intelligence? The Evidence Is Clear.” Nature 650, no. 8100 (2026): 36–40. https://doi.org/10.1038/d41586-026-00285-6.

This authors argue that AGI is no longer a future goal but a present reality. Written by a multidisciplinary team of experts, the piece posits that current large language models (LLMs) have already met the functional standards for general intelligence.

The authors advance the idea that a major hurdle in acknowledging AGI is how we define it. The authors argue that critics often set the bar too high, requiring "superintelligence" or "perfection" that even humans don't possess. They suggest that general intelligence should be judged by:

• Breadth: Competence across multiple domains (math, language, science).
• Depth: Strong performance within those domains.

By these standards, if we consider a person like Einstein to have general intelligence despite his inability to speak Mandarin, we must grant the same status to AI, which exhibits broad and deep cognitive flexibility.

I suggest that AI is a valuable tool for genealogical research and that fears and efforts to limit its use are merely the same responses that occurs with technological innovations generally. 

 For example, years ago with the introduction of the hand calculator, schools throughout the United States and perhaps the world immediately reacted that reliance on hand calculators would result in the destruction of mathematics, and that students would not learn because they had this mechanical crutch.

The same arguments were made about the Wikipedia website that it would ruin education, and that children in schools should not use the website because it was unreliable and would prevent them from learning real information (however that was defined).

Today we're seeing the same thing in the academic world with alarms being raised over the fact that students will not learn how to do basic skills such as recognizing old handwriting because they will somehow need that. The answer is, yes, AI can make mistakes and is limited in many ways, but it is a tool, not a replacement for human knowledge. It's necessary as a component in the use of AI, just as it is with driving cars, flying airplanes, and any other major activities in the world.

A good analogy would be to require automobile mechanics, working on today's advanced automobiles, to retain horseshoeing skills because fundamentally, that was one of the abilities that was necessary for a large number of people a hundred years ago. 

Genealogists, like the rest of the world, will have to adapt to AI just as they have to the rest of the world's technological advancements. Get on the train before it leaves the station.