AI for Genealogy: How to Use AI for DNA Genealogy - Analyze Genetic Matches & Break Through Brick Walls Safely

Staring at hundreds of DNA matches with no idea where to start? Wondering if AI can help solve your genealogy DNA mysteries? In this groundbreaking episode of Ancestors and Algorithms, discover how to safely combine DNA results with AI chatbots to revolutionize your family history research—without compromising anyone's privacy.

What You'll Learn in This Episode:

This comprehensive guide teaches genealogists exactly how to use AI tools like ChatGPT, Claude, Gemini, and Perplexity to analyze DNA matches, understand centimorgan relationships, and create research strategies that actually work.

DNA Basics for Beginners: Learn what centimorgans mean, how DNA matching works, and why that 387 cM mystery match might be your key to breaking through a brick wall. We break down relationship ranges from parent/child (3,500 cM) to third cousins (50-100 cM) in simple, understandable terms.

Privacy Protection First: Discover what you should NEVER share with AI chatbots (including raw DNA files and identifying information about living people), plus step-by-step instructions for opting out of training data on ChatGPT, Claude, Gemini, and Perplexity. Your family's genetic privacy matters—learn how to protect it.

AI Chatbot Comparison for DNA Work: Get the complete breakdown of which AI tools work best for DNA genealogy. Compare ChatGPT (with GPT-5 and Thinking models), Claude Sonnet 4.5, Gemini 2.5 Flash and Pro, and Perplexity's multi-model approach. Learn which features matter for genetic genealogy and whether free versions are enough.

5 Practical DNA + AI Techniques You Can Use Today:

• Relationship Analysis: Turn centimorgan numbers into actionable research leads
• Match Pattern Recognition: Identify how multiple matches connect to each other
• Research Strategy Planning: Get AI to suggest which records to search and why
• Creating Descendancy Charts: Visualize complex relationships with AI-generated diagrams
• Learning DNA Concepts: Master X-DNA inheritance, triangulation, and more

Real-World Case Study: Follow along as we walk through an actual DNA mystery solved using AI assistance. See exactly what prompts were used, how the AI responded, and how traditional genealogy research confirmed the findings. This isn't theory—it's practical application you can replicate.

Safe AI + DNA Integration: Learn what IS safe to share with AI chatbots: anonymized centimorgan numbers, match patterns without names, general DNA questions, and research strategy queries. Discover how to give AI the puzzle pieces it needs without compromising privacy.

Perfect For: Adoptees searching for biological families, genealogists with unknown ancestors, anyone overwhelmed by DNA match lists, researchers wanting to combine traditional methods with AI innovation, and family historians who want to work smarter, not harder.

Keywords: AI genealogy, DNA matches, genetic genealogy, ChatGPT for genealogy, centimorgans explained, DNA privacy, AI chatbots, ancestry research, family history AI, adoptee search, unknown ancestors, DNA analysis

Connect with Ancestors and Algorithms:

📧 Email: ancestorsandai@gmail.com
🌐 Website: https://ancestorsandai.com/
📘 Facebook Group: Ancestors and Algorithms: AI for Genealogy - www.facebook.com/groups/ancestorsandalgorithms/

Golden Rule Reminder: AI is your research assistant, not your researcher.

Join our Facebook group to share your AI genealogy breakthroughs, ask questions, and connect with fellow family historians who are embracing the future of genealogy research!

New episodes every Tuesday. Subscribe so you never miss the latest AI tools and techniques for family history research.

Transcript

You know that feeling when you're staring at a DNA match list with 47 second cousins, and you have absolutely no idea which one might lead you to your great-great-grandmother? Yeah, me too. What if I told you that A. I. chatbots could help you organize those matches, analyze relationship patterns, and even suggest which research path to take next, all without compromising anyone's privacy? I'm about to show you how to safely combine two of genealogy's most powerful tools, your DNA results and artificial intelligence. And trust me, once you see what's possible, you're going to wonder how you ever did DNA research without it. 

Welcome to Ancestors and Algorithms, where family history meets artificial intelligence. I'm your host, Brian, and this is episode 14, DNA plus A. I. If you've been following along with this podcast, you know we've been exploring how A. I. can help with everything from reading old handwriting to writing family biographies. But today? Today we're tackling something genealogists have been asking me about constantly. Can A. I. help with my DNA matches? The short answer is yes. The long answer? The one that's going to keep you and your living relatives safe while revolutionizing your research? That's what we're diving into today. Let me tell you about my friend Sarah. Now, Sarah is a solid genealogist. She's been researching her family for about eight years, has a beautiful family tree on ancestry, and she's really good at working with historical documents. But when she took a DNA test last year, she hit a wall. A big one. Sarah's grandmother was adopted in 1942, and Sarah has been trying to identify her biological family for years. When that DNA test came back, she had over 1,200 matches. 

And the closest match that wasn't a known relative was someone named DJK1978, who shared 387 centimorgans with her. Sarah did what most of us do. She stared at that number. 387 centimorgans. She googled it. She looked at the shared CM project chart. She knew this person was probably a second cousin, maybe a first cousin once removed, possibly a half-first cousin twice removed. There were like seven different possibilities. Then she looked at DJK's 1978 family tree. It had 12 people in it. No surnames in common with Sarah's tree. And DJK1978 hadn't logged into Ancestry in three years. Sarah called me, frustrated. I have all this DNA data, she said. But I don't know what to do with it. don't know which matches to focus on, how to organize them, or even where to start. And here's the thing. Sarah's situation is not unique. So many genealogists are sitting on DNA results that could crack their biggest brick walls. But they're drowning in data. This is where AI comes in. But not in the way you might think. Here's what I need you to understand right from the start. AI chatbots cannot access your DNA databases. They can't search Ancestry or 23andMe for you. They can't look up your matches or run comparisons in the background. But what they can do is help you think through your DNA puzzles, organize your information, analyze patterns, and suggest research strategies based on the relationship you're seeing. 

AI is your research assistant, not your researcher. That golden rule we talk about in every episode? It applies double when we're talking about DNA. Because DNA data is different. It's permanent. It's about you and your living relatives. And we need to handle it with care. So in this episode, I'm going to show you exactly how to safely use AI chatbots to supercharge your DNA genealogy work without putting anyone's privacy at risk. Alright, before we dive into the AI stuff, we need to make sure everyone's on the same page about DNA basics. If you're already a DNA pro, bear with me for a few minutes. But I promise this will be quick, and it'll set us up for the AI applications coming next. When you take a DNA test with companies like Ancestry, 23andMe, or MyHeritage, you get what's called autosomal DNA results. And the main thing you're looking at is your match list. All the people you share DNA with. The unit of measurement we use is called a symptom organ, abbreviated as C-Capital-M. Think of centimorgans like inches or centimeters, except instead of measuring physical distance, they're measuring genetic distance. It's a unit that tells us how much DNA you share with someone. Every person has about 6,800 to 7,200 centimorgans of DNA total. You got roughly 3,400 from your mom and 3,400 from your dad. When you share DNA with a match, the number of centimorgans you share tells us approximately how closely you're related. Here's a quick reference guide. You share about 3,500 cm with a parent or child. Full siblings share 2,500 to 2,900 cm. First cousins share 700 to 1,000 cm. Second cousins share 200 to 300 cm. Third cousins share 50 to 100 cm. Now, here's where it gets interesting for AI purposes. That centimorgan number is just a clue. For example, if you share 850 cm with someone, they could be a first cousin, a half-sibling, a great-aunt, or even a half-great-aunt. There are multiple possible relationships that fit that same amount of shared DNA. That's the puzzle genealogists are constantly solving. Taking that centimorgan number, looking at family trees, and figuring out the actual relationship. And this is exactly where AI can help. Okay, now let's talk about the most important part of this entire episode. Protecting I want to be very clear about what you should never share with AI chatbots. Never upload your raw DNA data file. That's the complete genetic file you download from Ancestry or 23andMe. It's usually labeled something like AncestryDNA underscore raw underscore data dot txt, or 23andMe underscore raw underscore data dot txt. That file contains your complete genetic code. It's permanent, it can't be changed, and it identifies not just you, but your entire family. Do not upload that to any AI chatbot. Period. Never share identifying information about living people without their consent. This means names, birthdates, current locations, anything that could identify your living relatives. Remember, your DNA reveals information about your parents, siblings, children, cousins, your whole family tree. When you share your DNA information, you're sharing information about them, too. Be cautious about training data policies. Most AI chatbots have an option to opt out of having your conversations used to train their models. Let me quickly run through where to find these settings. For ChatGPT, go to your profile, click Settings, then Data Controls, and toggle off Improve the Model for Everyone. For Claude, go to Settings, then Privacy, and toggle off Help Improve Claude. For Gemini, go to MyActivity. Google.com forward slash Product forward slash Gemini. And turn off Gemini Apps Activity. For Perplexity, in Settings, find Preferences, and then AI Data Retention, and toggle that off. Now, I'm not saying this to scare you. Now, I'm not saying this to scare you. I'm saying this to scare you. I'm saying this to empower you. Because when you know how to protect privacy, you can use these tools with confidence. So, what can we safely share with AI? A lot, actually. Centimorgan numbers without names. You can absolutely share things like, I have a match who shares 387 centimorgans with me. Or, match A shares 850 centimorgans. Match B shares 425 centimorgans. The AI doesn't need to know who these people are, just the relationships. Anonymized match lists. Create a simple spreadsheet or list where you label your matches as match 1, match 2, match 3, and include their centimorgan amounts and maybe what side of the family you think they're on. No names, no usernames, no identifying details. General DNA questions. You can ask things How are two people related if they share 245 centimorgans? Or, explain XDNA inheritance to me. Or, what does it mean when three people triangulate on chromosome 7? Research strategy questions. This is where it gets really powerful. can say, I have three second cousin matches on my paternal side, and they all descend from different children of the same couple. What research strategy should I use to identify my unknown great-great-grandfather? See how we're giving AI the puzzle without giving it anyone's private information? That's the sweet spot. Now, let's talk about which AI chatbots are actually available and what they can do. As of October 2025, you've got several solid options. ChatGPT from OpenAI has multiple versions. There's a free version, a Go version for about $5 a month in select countries, plus for $20 a month and Pro for $200 a month. For most genealogists, the free version or plus version will work just fine for DNA questions. ChatGPT has GPT-5, their flagship model, and also GPT-5 thinking for more complex reasoning tasks. Claude from Anthropic also has tiers, free, Pro at $20 a month and Max at $200 a month. Claude Sonnet 4.5 is their latest model and it also has a thinking version. I find Claude particularly good for organizing complex information and creating diagrams, which can be really helpful for DNA work. Gemini from Google has a free version and an advanced version at $20 a month. Gemini has two main models, 2.5 Flash, which is fast, and 2.5 Pro, which is more powerful. Gemini is great if you're already in the Google ecosystem. Perplexity is interesting because it gives you access to multiple AI models in one place. The free version works well, but Pro at $20 a month lets you choose between GPT-5, Claude, Gemini, Grok, and others. If you like testing different AI models for the same task, Perplexity is worth considering. For DNA genealogy work specifically, I've found that Claude is excellent for creating visual diagrams of relationships. ChatGPT is great for explaining complex concepts in simple terms, and Perplexity is fantastic when you need it to search current genealogy resources while thinking through your problem. You don't need to subscribe to all of them. Start with the free versions, see which one you vibe with, and then consider upgrading if you find yourself using it daily. Alright, now for the good stuff. Let me show you five specific ways to use AI with your DNA results. Technique number one, relationship analysis. Create a simple list of your top unidentified matches with their centimorgan numbers. Then ask AI, Quote, 

The

AI will give you probability ranges and explain why certain relationships are more likely than others. Technique number two, match pattern recognition. Let's say you have six matches who all seem to be related to each other, but you can't figure out the pattern. Create anonymized Match one shares 425 centimorgans with me. Match two shares 380 centimorgans with me and 850 centimorgans with match one. Match three shares 200 centimorgans with me and 425 centimorgans with match one. Then ask, Quote, Based on these shared centimorgans amounts, how are these people likely related to each other and to me? End quote. Technique number three, research strategy planning. This is where AI really shines. You can say, Quote, I'm trying to identify my great-great-grandfather born around 1880 in Ohio. I have three second-cousin DNA matches who all descend from siblings in that generation, but their trees don't overlap with mine. What research strategy should I use to connect our families? End quote. The AI will suggest approaches like looking for siblings' marriages, checking migration patterns, or researching in specific record types. Technique number four, creating descendancy charts. You can use AI, particularly Claude, to help you create visual charts of your match's relationships. You can provide a family structure in text form and ask it to create a diagram showing how everyone connects. This is incredibly helpful for understanding complex relationship patterns. Technique number five, learning DNA concepts. Don't understand what XDNA is? Confused about how triangulation works? Ask your AI chatbot. Say, Quote, Explain XDNA inheritance in simple terms with an example. End quote. Or, Quote, I don't understand DNA triangulation. Can you explain it using a simple family scenario? End quote. And remember, this is where our golden rule comes back. AI is your research assistant, not your researcher. The AI is helping you think through the puzzle, but you're still the one who needs to verify everything with actual genealogical evidence. Let me walk you through a real example of how this works. Well, not real, I've changed details to protect privacy, but this is based on an actual situation I helped someone work through using AI. Meet Jane. Jane has been stuck trying to identify her great-grandmother's parents. She took a DNA test and got her results and she's been staring at her match list for weeks. Here's what Jane did. Step 1, she organized her top mystery matches. Jane created a simple document that looked like this. Mystery Match 1, 421 Centimorgan Shared. Mystery Match 2, 387 Centimorgan Shared. Mystery Match 3, 203 Centimorgan Shared. Mystery Match 4, 198 Centimorgan Shared. She noticed that Mystery Match 1 and 2 both had ancestors from Kentucky in their trees and Mystery Match 3 and 4 both had ancestors from Tennessee. Step 2, she asked Claude for help analyzing the pattern. Jane opened Claude and wrote, quote, I'm trying to identify my great-grandmother who was born around 1895 in Kentucky. I have four DNA matches I can't connect to my tree. Match 1 shares 421 Centimorgan. Match 2 shares 387 Centimorgan. Match 3 shares 203 Centimorgan. Match 4 shares 198 Centimorgan. Matches 1 and 2 both have Kentucky ancestors, while Matches 3 and 4 have Tennessee ancestors. What might this pattern tell me, and what's the most likely relationship scenario? End quote. Claude responded with an analysis suggesting that Matches 1 and 2 were probably second cousins, potentially from one side of her great-grandmother's family, while Matches 3 and 4 were likely second cousins once removed, possibly from the other side. It also suggested that geographic split might indicate her great-grandmother's parents came from different states. Step 3, she asked for a research strategy. Jane then asked, quote, given this pattern, what research strategy should I use to identify my great-grandmother's parents? What records should I prioritize? End quote. The AI suggested, research marriages between Kentucky and Tennessee families in the 1890s. Look for migration patterns from Tennessee to Kentucky. Check for land records showing families from both states and neighboring counties. Build out the trees of Matches 1 and 2 to find their common ancestors. Look for the surname patterns that appear in both Kentucky and Tennessee Matches. Step 4, she asked Claude to create a visual. Jane said, quote, can you create a simple diagram showing how these matches might connect to my unknown great-grandmother? End quote. Claude created a text-based family tree diagram showing the hypothetical relationships, which helped Jane visualize what she was looking for. Step 5, she verified with actual research. And here's the critical part. Jane didn't stop at AI. She took the AI suggestions and went to work in the actual historical records. She found a Kentucky marriage record, located census records, and confirmed the relationship through documentation. The AI had given her the road map, but Jane did the genealogy. That's the power of combining AI and DNA. The AI helps you see patterns, organize your thinking, and plan your strategy. But you're still the researcher who does the actual verification. Okay, it's time for your homework assignment. This week, I want you to pick one DNA mystery match. Someone you haven't been able to connect to your tree. Create an anonymized description. The symptom organ amount you share, any geographic patterns you notice, and what you know about their family tree without using names. open up a free AI chat bot, ChatGPT, Claude, or Gemini, and Perplexity, and ask it to help you analyze the relationship and suggest research strategies. Take notes on what the AI suggests. Then, and this is important, come to our Facebook group, Ancestors and Algorithms AI for Genealogy, and share your experience. We just hit 400 members this week. It's pretty exciting. And then tell us, what did you ask? What did the AI suggest? Did it help you think about your match in a new way? I want to hear your stories because I guarantee some of you are going to have breakthrough moments this week. And speaking of the community, I got an email last week from Lisa in Oregon who used one of our AI prompting techniques to finally identify her third great-grandfather. Lisa, that is amazing. I love hearing about your successes. If you have a DNA plus AI story to share, email me at ancestorsandai at gmail.com. I read every email, and I love featuring listener success stories. Now, one question I've been getting a lot. What about DNA Painter and the Shared Centimorgan Project? Great question. DNA Painter is a fantastic free tool created by genealogist Johnny Pearl and the Shared Centimorgan Project specifically helps you understand what relationships are possible for a given Centimorgan amount. It's completely free to use at dnapainter.com. You can actually use AI and DNA Painter together. Use DNA Painter to get the probability ranges for your match's relationship, then feed that information into AI and ask for research strategies for each possible relationship. They complement each other beautifully. Alright, before we wrap up, let me remind you, DNA plus AI isn't scary. It's not dangerous. It's just a new tool in your genealogy toolkit. And like any tool, when you use it properly, with care, with ethics, and with privacy protection, it becomes incredibly powerful. You're not replacing traditional genealogy. You're enhancing it. You're giving yourself a research partner that can help you think through complex patterns and plan better strategies. And that's pretty revolutionary. So here's what we covered today. First, we talked about DNA basics, what Centimorgan's are, and how relationships work. Then we established the Privacy Foundation. What never to share, and what's safe. We explored which AI chatbots are available and what makes each one useful. And finally, we walked through practical techniques for combining AI and DNA safely. Remember our golden rule, friends. AI is your research assistant, not your researcher. Let it help you organize, analyze, and strategize. But you're still the genealogist who verifies everything with real evidence. If you found this episode helpful, please hit that follow button wherever you listen to podcasts. Leave a rating and review. It helps other genealogists find the show. And join our Facebook group, Ancestors and Algorithms AI for Genealogy, where we're building a community of researchers who are using this tool ethically and effectively. This has been Ancestors and Algorithms. I'm your host, Brian. And remember, your ancestor stories are waiting to be discovered. AI is just helping you find them faster. Now get out there and solve some DNA mysteries. Let's go.