By Nancy Richey (Special Collections Librarian/Professor, Western Kentucky University Libraries)
Column Editor: Lesley Rice Montgomery, MLIS (Catalog Librarian II, Tulane University Libraries’ Technical Services Department)
Against the Grain V38#3
Introduction
Ms. Richey serves as Special Formats Librarian for the Department of Library Collections & Discovery at Western Kentucky University, where she has been a faculty member since 2008. She earned degrees in Information Science and Southern History from both the University of Kentucky and WKU.
Her work focuses on collection development, archival stewardship, research support, and instructional outreach within Special Collections. She provides research assistance and library instruction for students, faculty, and community researchers, while also leading efforts to expand and preserve the Department’s distinctive image and historical collections.
Ms. Richey has served on numerous historical and preservation organizations, including the Janice Holt Giles Society, the Morrison Park Camp Meeting Site Restoration Board, and the Daughters of the American Revolution. She is the author of two local history volumes in the Images of America series and co-author of two biographical works centered on noted Kentuckians.
Ms. Richey is joining us this month to share her valuable expertise and personal insights as she discusses the use of AI technologies when conducting genealogical research. — Lesley
Using AI Technology in Genealogical Research
As an academic librarian and experienced genealogist, I am excited about the current and future possibilities of the use of AI in genealogical research and I want to be able to assist a diversity of patrons with their traditional and now, genealogical AI questions. I, like many librarians, am excited at the increased access to records for previously under-documented groups. At present, major platforms such as Ancestry and Ancestry Library Edition (ALE) have over twenty-five million users, and the genealogy industry is projected to have a growth rate of over 11% annually.1 I am continually surprised at its evolution and genealogical possibilities. The use of AI in genealogical research is not new; popular commercial genealogy websites such as FamilySearch and Ancestry have been using algorithms behind the scenes since the 2000s. But for the individual, AI offers the promise of streamlined research, enhanced accessibility, and multimodal analysis. Its capabilities include transcription and multilingual translation, photo creation and enhancement, document analysis, and DNA interpretation. This article is designed to help any individual or librarian who may be in the novice stage but who wishes to understand terms and uses of the major AI models. I will explore some of the well-known sites and will provide short reviews. As an information professional I do not want to become a “dinosaur,” but I want to keep up with current research methods and tools to better assist my patrons and to sharpen my own research skills. As I think of this, I like to remember the following dialogue from the 1956 movie, The Invasion of the Body Snatchers:
Dr. Miles J. Bennell: I don’t want any part of it.
Dr. Dan ‘Danny’ Kauffman: You’re forgetting something, Miles.
Dr. Bennell: What’s that?
Dr. Kauffman: You have no choice.
ChatGPT — https://chatgpt.com/ — Chat Generative Pre-trained Transformer, (OpenAI) is a great place to start in learning the uses of AI in genealogy. It can help with writing, brainstorming, and creating narratives from your research. It can also create images (an extra fee) and analyze data. The chatbot, developed by OpenAI, was launched in 2022. Its user-friendly search box greets you, “What Can I Help With.” Visually it is uncluttered and offers its own directions “How To Use ChatGPT” in the search bar. You must agree to their terms and privacy policy. Remember this is an LLM (large language model), and it is constantly harvesting information, your information. You cannot put in a prompt such as “give me the name of my third great-grandmother, or I want the life history of Thomas Richey born 1802 in Allen County, Kentucky.” You will quickly receive a polite answer, “I’m sorry, but as an AI language model, I don’t have access to personal information or historical records that could help me to provide specific information about individuals. However, here are some general tips and resources that could help you in your research.” ChatGPT has a free component with usage limits on messaging and file uploads. The cost for the “Plus” version is $20/month.
So, here are some good genealogical prompts for these sites: Can you help me interpret the text of a family document? Can you analyze surname origins and variations? What are some online databases where I can find historical records relevant to my family history? Can you help me contextualize the life of an ancestor within historical events and societal trends? I have a lot of genealogical data. Can you help me organize the data into a timeline? Can you help me craft a story about a family member with the genealogy records I have?2
My father was a great storyteller and passed on many interesting family tales, so I wanted to prompt ChatGPT to create a narrative for me. One story my father told me involved the assassination of one of my great-uncles (Walter “Bud” Richey) in Allen County, Kentucky, in 1892. I had snippets of information from my father, and from extant newspaper articles that included the limited facts of the murder. I entered into ChatGPT some of the following types of facts and keywords: Bud Ritchey, assassination, Moore, Eliza, his wife, William Moore suspected, wealthy, bullet and throat and uploaded the newspaper story with the most information. By entering these facts found in newspaper stories and other limited sources, I prompted ChatGPT to create a story from this limited input. Below is the exciting and melodramatic story that was generated.
“In the rolling hills of Allen County, Kentucky, in the autumn of 1892, the Ritchey family’s story took a tragic turn that would echo through generations. Bud Ritchey, a man known for his wealth, his land, and his penchant for trouble, met his end in a cowardly ambush on a public road near Gainesville. It was the third attempt on his life, and this time, the assassin succeeded. On that fateful afternoon, as the sun dipped low in the sky, Bud and his wife Eliza were traveling along the road, unaware of the danger that lurked nearby. Suddenly, shots rang out, and Bud fell, a bullet tearing through his neck and windpipe. Eliza, horrified, witnessed the brutal assassination of her husband. William Moore was suspected, a man with whom Bud had a long-standing feud. This was the third attempt on Bud’s life, but this time, the outcome was fatal. The assailant was never charged, and the Ritchey family was left shattered and heartbroken. Bud Ritchey was a well-known figure in the area, a wealthy stock trader and landowner. His overbearing and quarrelsome nature had made him enemies, but nothing could have prepared his family for such a senseless act of violence. The aftermath was grim. Bud’s body was so badly mutilated that he was unrecognizable. His house was later burned, adding to the family’s grief and confusion. In the end, the rest of the Ritchey family made the difficult decision to leave Allen County behind, seeking solace in Barren County. But the story did not end there. Bud’s wife Eliza, devastated by the loss of her husband, lived to be 93-years-old but never remarried, her heart forever haunted by the memory of that tragic day. Their daughter, in a twist of fate, married the nephew of Bud’s assassin, a cruel reminder of the tangled web of life and death. Today, Bud Ritchey rests in peace in the Richey Graveyard, a somber reminder of a man whose life was cut short by cold-blooded murder. His legacy lives on in the stories passed down through the generations, a testament to the resilience of a family torn apart by tragedy.”
In addition, I wanted an image to go along with the story. ChatGPT’s Dall-E image generator created the below image with prompts from the event.
The remembered story really came alive for me through the narrative and the photo created by ChatGPT’s Dall-E.
Some Major Sites to Consider
Claude (Anthropic) — https://claude.ai/login is excellent at handling long genealogical documents, analyzing, summarizing, and even transcribing texts. With effective, iterative, and clear prompts, Claude can process large amounts of data and draft cogent narratives as well as transform raw data from spreadsheets and research logs into more usable and sharable information. One researcher noted Claude “when asked to analyze my genealogy research log and provide a comprehensive summary,” quickly “delivered a solid summary with insights.”3 Claude can also create a family history website and help to interpret data dealing with genealogical brick walls. Claude also stands out because of the high-quality accurate answers. Claude has a limited free use that can write, edit, and create content, but with the Pro version ($17/month billed annually), you can analyze more text and images, search the web, use extended thinking for more complex work, and connect to Google.
Gemini (Google) — https://gemini.google.com/app is called “your personal AI assistant” by the site and lists tools that create images, video, code, and other media, and audio overviews. Gemini offers good verification and helps with workflows and organizational tasks. For Gemini, and all AI tools, the use of detailed, request-specific prompts provides the best results. Since this is a Google product, it feels very familiar. It performs a “double-check” response linking directly to a Google site providing the sources of the information and has access to YouTube and maps. It has the longest context window of any AI model, allowing for more input and a reduced need for retraining. Gemini has an excellent free component with Imagen 4, Veo, and 15 GB storage. The pro version is $20/month.
Copilot (Microsoft) — https://copilot.microsoft.com/ has OneDrive/file Searching capabilities so that the model can “search across all files such as Word, Excel, PowerPoint, Outlook, OneNote, Whiteboard, SharePoint and Teams to find connections between different documents.” Additionally, it “can be used to manage and summarize meeting notes or interview recordings, which is helpful when interviewing living relatives.”4 One fun prompt is to ask that a photo be described in narrative form with scene, setting, people, architectural details, activities, and historical context. It has a read aloud feature, and you can also talk to Copilot with Copilot voice and use Copilot vision to respond based on what websites or other items you are viewing. I like the fact that its responses are to-the-point and can even provide a sample picture of something in the prompt. It also sets its limitations in that it quickly surrenders, saying it does not have enough information to answer the prompt. Copilot has a free version with limited features but offers Copilot Pro at $20/month.
Perplexity AI (privately owned) — https://askgpt.app/ai/perplexity is an AI search engine that provides cited answers and is always directly connected to the internet. In the Pro Search, the bot can ask follow-up questions that prompt more research on the part of the user. Since footnotes are listed for a search, I can verify information for myself. By using the Pro version, I can also choose various LLMs (large language models) and even in-house models. As well, I can focus on specific sites to search, giving me a more personalized feel for my queries. Related topics, examples, and other sites may be suggested. Multimodal capabilities are limited. Like many of the sites, it finds information about records, but it may not find the records. Perplexity is free in its basic version. The Pro version is $20/month.
Other Genealogical / Family History sites using AI
Kindex (privately owned app) — https://app.kindex.org/ is a cloud-based archive platform that allows family historians “to digitize, transcribe, index, and search private family records, letters, and documents. It allows a user to create searchable archives, make documents like handwritten journals accessible rather than just storing images, [and uses] AI models for transcription and image captioning.”5 This is a great collaborative site.
FamilySearch Full-Text Search (Church of Jesus Christ of Latter-day Saints) — https://www.familysearch.org/en/united-states/ is a remarkable free site that uses AI to instantly search millions of historical records that have been scanned. This full-text search feature enables you to use keywords and surnames to see scanned handwritten and printed relevant documents. This “approach also supports multilingual and cursive-script recognition, broadening accessibility to global record collections.”6
Deep Nostalgia AI (D-ID) — https://deep-nostalgia-ai.com/ is a one-of-a-kind site that allows you to animate the subjects in your old family photos, creating living portraits or animated video. The results can be startling and eerie.
Problems with AI in Genealogy
The uses of AI in genealogical research are many. Additionally, as an academic librarian, I have used the tools in other ways such as creating subject lists for manuscript collections and oral interviews. What a timesaver! However, I am learning that along with the many promises of AI come many concerns. Sometimes, the bot wants to provide as much genealogical information as it can, so it is redundant or it simply hallucinates or makes up information. A researcher noted “that’s part of the problem — it’s an LLM (large language model), meaning it guesses at the next most likely word or character rather than being actual research.”7
Additionally, there are privacy issues, misuse of data, bias, false information, hallucinations, and confabulations. Moreover, issues of copyright, lack of transparency of sources, and the high cost involved in accessing some of the information are failings. A genealogist observed “AI produces language, not verified truth. It generates what sounds right — not what is right. Incorrect information is delivered with the exact same confidence as correct information.”8
Researchers are also looking closely at emergent abilities and what that may mean for the future.9 There are many misuse stories out there (see AI Incident Database10) such as a lawyer using ChatGPT to write a brief, only to have it create fake legal citations. In using AI for genealogical research or any research, it is up to you and me to learn about these tools, their strengths, and limits. The Coalition for Responsible AI in Genealogy offers to us five guiding principles to ensure genealogical integrity: accuracy, disclosure, privacy, education, and compliance (ADPEC).11 Try to go back to the original source, if possible and apply recognized genealogical standards of evidence. Also, remember your information input is being harvested for training purposes; read user agreements carefully. I have learned that there is no substitute for professional judgment and the human touch, but what an exciting adventure this new genealogical AI world is going to be. Oh, the Places You’ll Go!
References and Further Reading
Kennedy, M. (2025, May 20). Ethics and Best Practice of AI Use in Genealogy Research. https://genealogy.org.nz/Ethics--Best-Practice-of-AI/11482/
Salazar, G. T. (2025, January 24). Unraveling the Past: AI and the Evolution of Genealogy. Association for Women in Science (AWIS). https://awis.org/resource/unraveling-the-past-ai-and-the-evolution-of-genealogy/
Schifferdecker, K., & Richey, N. (2024). Artificial Intelligence and Genealogy: Promise and Problems. Kentucky Libraries, 88(3), 1–16.
Tanner, A. (2024, August 15). AI Developments in Genealogy and How they Impact You. https://www.familysearch.org/en/blog/ai-developments-genealogy
Endnotes
1. Genealogy Products and Services Market Report for M&A, Expansion, and Competitive Benchmarking (2025–2034) (2025, June 25). In EIN Presswire. https://www.einpresswire.com/article/825455670/genealogy-products-and-services-market-report-for-m-a-expansion-and-competitive-benchmarking-2025-2034?utm_source=chatgpt.com
2. Tanner, A. (2026, February 2026). https://www.familysearch.org/en/blog/ai-developments-genealogy
3. Allen, D. (2025, August 27). I tested Claude, ChatGPT, and Gemini on my research notes. Here’s who won. https://denyseallen.substack.com/p/ai-genealogy-tools-comparison
4. Dunn, M. (2026, February 27). Is Microsoft 365 Copilot Actually Useful? We Put It to the Problem Test. https://www.youtube.com/watch?v=jxhifAeoTiI
5. Lee, D. N. Amazing AI Tools for Genealogy In One Place – Kindex Archives. https://www.familyhistoryfanatics.com/kindex-ai-tools-for-genealogy
6. https://www.familysearch.org/en/search/full-text/
7. https://www.legacytree.com/blog/using-ai-for-genealogy-research
8. Little, S. (2026, March 5). https://www.familysearch.org/en/rootstech/session/ai-family-history-foundations-first-steps-mastering-ai-basics-for-genealogy
9. Ornes, S. (2024, March 24). Wired. https://www.wired.com/story/how-quickly-do-large-language-models-learn-unexpected-skills/
10. https://incidentdatabase.ai/
11. Guiding Principles for Responsible AI in Genealogy. https://craigen.org/
Thank you for sharing this incredibly detailed and digestible overview of all of the most recent AI tools and updates that we can utilize as genealogists. I am sharing your article with my fellow Knickerbocker DAR Board Members.
Sincerely,
Nicole Ameduri, Chaplain of the Knickerbocker DAR Chapter
Mayflower Society