Don't get left behind by AI

 

Artificial intelligence is not a passing fad. It is a fundamental change affect nearly every aspect of our lives. Genealogy does not escape the effects of these major changes. One obvious effect of AI to FamilySearch.org is the end of volunteer-based indexing. If you haven't noticed, Indexing has disappeared from the website. 

All the indexing of records on the FamilySearch.org website is being done by AI-driven Full-Text Search based, in part, on AI based handwriting recognition. 

FamilySearch.org/labs

I suggest the following to get some idea about what is going on. 

• CRAIGEN.org The website of the Coalition for Responsible Use of AI in Genealogy. 
• https://youtu.be/Hfwxe7llx7Q?si=EMEV7WzXMaKxulu6  Updated-AI, Handwriting Recognition and Full-Text Research on FamilySearch

I have a number of other videos on the BYU Library Family History Center YouTube Channel,  but they are now months old and possibly already far out of date. This is the same problem with looking online for current information about AI. You have to look at the date of the article or video and try to find something that is not already out of date. 

AI is currently reaching the saturation point of online text and videos. By the way, I have more videos about AI coming up during the months ahead. 

FamilySearch Improved Merge Experience

 

One basic fact about online family trees is their propensity to create duplicates. This apparently stems from the belief by each family tree contributor that their work is somehow unique and completely accurate. When this belief in uniqueness and accuracy is extended to the large online family tree websites, the results is the larger the family tree the more likely the possibility of duplication. A huge, collaborative, universal family tree such as the FamilySearch.org Family Tree is by its collaborative nature to provide a place for all those believing in uniqueness and accuracy. The number of duplicates entries on the FamilySearch.org Family Tree has been, in the not too distant past, into the hundreds of millions. According to FamilySearch.org Facts, there are about 1.72 billion people in the Family Tree. Even if only 1% of that number were duplicates, there would be about 17.2 million duplicate entries. So, what is the actual percentage?

The actual number of duplicates could only be determined by some sampling method that took large numbers of people in the Family Tree and looked at them to see if they had any duplicate entries. Currently, in some cases, the Family Tree will tell users that there are duplicate entries. But, in my own experience, some duplicates only show up when information is added to one of the entries. For example, if I have an entry for a person with a very common name, when I do the research, I often see duplicates show up and if I am working with a family, those duplicates can continue with all the family members. This happens frequently enough for me to assume that there are a huge number of duplicates in the Family Tree. 

The Family Tree has several features that attempt to prevent duplication. However, a new person can be added to the Family Tree simply by indicating that there are no matches to existing entries. When the new entries are being added by someone uploading a GEDCOM file or working on a linking project duplicate entries are almost guaranteed because the standard FamilySearch method for determining duplicates does not find duplicates that are only discovered through doing research by adding sources and verifying all the information. 

Here is an example of a person with an extremely common name that has been been added with no sources from a GEDCOM file. 

There is obviously not enough information about this person to determine if there is a duplicate entry. Her husband is Marcos Garcia de Jove PSG1-PCF. He also has no sources. The issue here is that by doing more research on this person who has no birth date and no death date finding duplicate entries is fairly common. That is true here because of the huge number of people named Maria Velazquez and also the fact that we do not have her full name and any information about her birth, marriage or death. 

Here are some recent articles about the new Merge Experience. 

• Grant, Kathryn. “FamiilySearch Family Tree Merge Experience Has a 2025 Update • FamilySearch.” FamilySearch, July 11, 2025. https://www.familysearch.org/en/blog/family-tree-merge-2025-update. 

• Id: 935, Article. “Merge Duplicate Records in Family Tree • FamilySearch.” HLPC Help Center, April 20, 2020. https://www.familysearch.org/en/help/helpcenter/article/how-do-i-merge-possible-duplicates-in-family-tree. 

• Tennant, Amie. “How to Fix Incorrect Merges in Your Family Tree • FamilySearch.” FamilySearch, March 10, 2025. https://www.familysearch.org/en/blog/fix-incorrect-merges-on-familysearch.

If you have questions about the new Merge, you can get help from the FamilySearch.org Community. 

https://community.familysearch.org/en/

FamilySearch Computer Generated Family Trees are not Sources

 

https://www.familysearch.org/ark:/61903/2:5:7ZXY-254?lang=en

In working with another missionary/volunteer at the BYU Library Family History Center, we ran across this source entry.

You can click on the image to enlarge it. This entry was apparently posted by FamilySearch.org. Note that there are 35 sources for this individual and this Computer Generated Tree was posted on April 23, 2024. The same day, FamilySearch.org added a copy of a parish register enty. 

The first entry, the Computer Generated Trees entry, is not a source. It has duplicate names for an existing record and an unsupported duplicate entry for the existing individual and his parents. The second entry, the parish register, is one that has the valid historical source information. The Computer Generated Trees entry come for this file in the Genealogies section of the FamilySearch.org website. 

This entry must be one of a multitude of entries being made by FamilySearch, apparently, all of them or many of them duplicates of entries for a person already in the Family Tree. Incidentally, the parish register is also a duplicate of an entry made in 2019. 

Although the specter of Computer Generated Trees is being discussed as a replacement for human intervention in the genealogical research process, this single entry seems to indicate that at least back in 2024, the process did nothing more than duplicate the work already done. I suppose that you could take the position that people are really superfluous and that computers can do a better job of building family trees given the access to the necessary information but from what is shown here, the reality would be a duplication on a massive scale. 

Are you ready to be replaced? I don't think I am either. 

The Amazing Escapades of William Francis Tanner L19W-7W9

 

https://www.familysearch.org/en/tree/person/details/L19W-7W9

Remember, as this story unfolds, that "Important research has been done on this person. Please read these alert notes before making changes." Also, he has a hefty 60 sources attached. Let's look at this amazing person. Oh, I should mention that the information in the FamilySearch Family Tree changes daily. I should also mention my research in Rhode Island has shown multiple people with the same William Tanner name during the years from 1657 to 1740. 

Here is his wondrous story. 

William Francis Tanner was suposedly born on 10 March 1657 in Chipstead, Surrey, England. There is a note that first mention of William in Rhode Island was in 1682 when he is 25 years old. However, he is married to Elizabeth Colgrove GWPP-Y78 who was apparently born in Sussex, England in 1665 and died in Sussex after the couple had five children who are born from 1680 to 1700 also in Sussex. Meanwhile, William was busy in Rhode Island by 1682. But he has to return to Sussex for children born in 1684, 1689, 1690, and 1700. 

In Rhode Island, William marries Hannah Tibbetts L6PR-VD4 in 1690 so he crosses the Atlantic just in time to have his daughter in Sussex in 1690. Quite a world traveler. He only manages to have one child with Hannah who dies in 1688. He quickly marries another woman, Mary Babcock, in 1690  and fourteen additional children born in 1691 (back to Rhode Island), two children in 1692, 1694, 1698, 1700, 1704, 1710, 1712, 1714, 1717, 1719, Deceased, and Deceased. Meanwhile, remember those five children born in England. He has multiple trips across the ocean. 

This William Tanner is currently only related to me through Sir Richard Vernon G4BD-PSP born in Derbyshire, England in 1415 and who died in 1451. 

This prolific William Tanner started out as my direct line ancestor before he was taken over by those Family Tree users who don't seem to have anything productive to do except make changes to this unfortunate family. My direct line is through another William Tanner that does not have the problem of multiple ocean crossings and three wives. 

These Revolving Door people are a significant waste of time and give the Famiily Tree a bad reputation among responsible genealogists. By the way, there do not appear to be any sources that have William Tanner with a middle name in Rhode Island. 

https://blog.myheritage.com/2025/07/significant-enhancements-to-mystories/

Quoting from the MyHeritage Blog post "Introducing MyStories: A New Service to Turn Cherished Memories into a Printed Keepsake Book"

For 21 years, MyHeritage has inspired millions of people worldwide to discover more about who they are and where they belong. While family trees, historical records, and photos are integral to family history — the stories, memories, and anecdotes that form the essence of life often go undocumented. MyStories encourages people to capture their stories and preserve them for future generations in high-quality hardcover books, ensuring their stories are cherished for posterity.For 21 years, MyHeritage has inspired millions of people worldwide to discover more about who they are and where they belong. While family trees, historical records, and photos are integral to family history — the stories, memories, and anecdotes that form the essence of life often go undocumented. MyStories encourages people to capture their stories and preserve them for future generations in high-quality hardcover books, ensuring their stories are cherished for posterity.

Now, MyHeritage has added a new level of convenience and utility. Here is the quote from the blog post entitled, "New Features on MyStories: Audio Recording and Enhanced Editing."

Many people are eager to tell their stories, but some are more comfortable talking than typing. With the addition of audio recording and transcription, you can record your story in your own voice with the click of a button. Once recorded, MyStories uses AI voice-to-text technology to automatically transcribe the story. The transcribed story appears directly in the story editor, where you can edit the text and layout, and add relevant photos. The audio recordings are saved so you can access them later on and download them, if you wish.

I am trying to write one subject a week for this entire year. I will certainly put images in the sections I have already written and in the subsequent sections.  

Apple study reveals the limitations of current AI Large Reasoning Models

 

Apple Machine Learning Research. “The Illusion of Thinking: Understanding the Strengths and Limitations of Reasoning Models via the Lens of Problem Complexity.” Accessed June 23, 2025. https://machinelearning.apple.com/research/illusion-of-thinking.

It is about time that someone or some company did studies to determine the limitation of the currently popular Large Languag Models (LLMs) and AI Large Reasoning Models (LRMs). First, short explanation about LLMs and LRM.

A Large Language Model (LLM) is a type of artificial intelligence (AI) program designed to understand, generate, and manipulate human language. LLMs are characterized by their massive scale—they are trained on enormous datasets of text and code, often comprising billions or even trillions of words. (Google Gemini search)

also: 

A Large Reasoning Model (LRM) is an advanced type of artificial intelligence model that extends the capabilities of Large Language Models (LLMs) by specifically focusing on and enhancing their ability to perform multi-step logical reasoning and problem-solving. (Google Gemini search)

Quoting from an article posted by The Guardian (Marcus, Gary. “When Billion-Dollar AIs Break down over Puzzles a Child Can Do, It’s Time to Rethink the Hype.” The Guardian, June 10, 2025, sec. Opinion. https://www.theguardian.com/commentisfree/2025/jun/10/billion-dollar-ai-puzzle-break-down.)

Apple did this by showing that leading models such as ChatGPT, Claude and Deepseek may “look smart – but when complexity rises, they collapse”. In short, these models are very good at a kind of pattern recognition, but often fail when they encounter novelty that forces them beyond the limits of their training, despite being, as the paper notes, “explicitly designed for reasoning tasks”.

Apple used a child's game called "The Tower of Hanoi" that is relatively easy to solve to show that the reasoning ability of the major LLMs and LRMs fail when faced with challenges that can be solved by young children. 

My own experience using AI for a variety of purposes (such as the ones above) indicate that the advancements for historical/genealogical research are as limited as I expected. However, when we talk about handwriting recognition and the related progress is handling large numbers of documents will fundamentally change the way genealogists do their research.

The Digitization of the Vatican Library

https://digi.vatlib.it/

 Yes, you are reading right. There is a website with the current digitization of the Vatican Library. The images are watermarked but that does not interfere with doing research. From the website: "The project aims to digitize the entire Library' collection namuscripts : 80,000 (excluding the archival units) mostly from Middle Ages and Humanistic period. See https://digi.vatlib.it/news/#news-1 

Here is a description from Google Gemini of the digital library. 

The Vatican Library has embarked on an ambitious and ongoing project to digitize its immense and historically significant collections, making them accessible to a global audience. This initiative, primarily housed on the "DigiVatLib" platform, offers free access to a vast array of materials, transforming the experience of engaging with these irreplaceable artifacts.

The digital collections are remarkably diverse, encompassing a wide spectrum of historical and cultural heritage. Chief among them are tens of thousands of precious manuscripts, many dating back to the Middle Ages and Renaissance. These include not only theological and ecclesiastical texts but also classical Greek and Latin works, Arabic, Hebrew, and Ethiopic manuscripts, and even early copies of works by influential figures such as Aristotle, Dante, Euclid, Homer, and Virgil. The digitization process captures these intricate documents with high resolution, preserving their delicate details, illuminations, and sometimes even their original blemishes.

Beyond manuscripts, the digital library also features a substantial collection of incunabula (books printed before 1501), offering insights into the early history of printing. Furthermore, researchers and enthusiasts can explore archival materials and inventories, providing a window into the administrative and historical records of the Church. The collections are enriched by graphic materials, coins and medals, and various other printed materials that form special projects within the library's holdings.

The digitization effort serves a dual purpose: long-term preservation of these fragile and invaluable items for future generations, and opening up their contents to a much wider public. The platform utilizes International Image Interoperability Framework (IIIF) technology, enhancing accessibility and enabling scholars to easily compare and utilize these resources alongside those from other major institutions. While the project is ongoing and a significant portion of the vast physical collections is yet to be digitized, the Vatican Library's digital endeavors represent a monumental step in democratizing access to some of the world's most treasured intellectual and artistic heritage.