Open Channels FM
Open Channels FM
Balancing Tradition and Innovation in Genealogy During the Era of Artificial Intelligence
Loading
/

In this episode of Open Web Conversations, Bob Dunn continues with the new series “AI for the Rest of Us” by exploring how artificial intelligence is transforming fields outside of tech with a deep dive into genealogy.

Our guest, professional genealogist Charleen Mullenweg, sharing her years of expertise tracing family histories, discusses the evolving role of AI in research, and reveals how automation helps handle everything from sifting through vast historical records to filling out complex applications.

Tune in for a conversation about the surprising ways technology is reshaping the search for family roots, the ongoing need for human oversight and empathy, and what the future holds for genealogists in an AI-driven world.

Since 2005, Automattic has built tools for the open web including WordPress.com, WooCommerce, and Jetpack that are used by millions of people to create, sell, and publish online. They believe in ownership, flexibility, and open source and we’re grateful for their support. Learn more at automattic.com.


InMotion Hosting brings over 25 years of experience, NVMe-powered speed, and 99.99% uptime to every plan they offer. When you need help, you get a real human, not a bot, and they’ll migrate your site for free. We’re happy to have them in our corner supporting the conversations we have here at Open Channels FM. Find your plan at inmotionhosting.com


Logo of Omnisend featuring a stylized 'i' icon and the brand name in lowercase letters.

Omnisend just dropped SMS pricing to $0.007, and their migration team moves your automations, templates and contacts in five days, free. That means you could be saving up to 35% in less than a week. We’re glad to have them supporting the show and the community we’re building around it. Use the code OpenChannels and get 30% off your first 3 months of any paid plan.

Takeaways

  • AI is Transforming Genealogy Research: AI tools have dramatically reduced the time needed to sift through massive amounts of genealogical documents and records, allowing professionals like Charleen to focus more on in-depth research and less on repetitive tasks (06:17, 12:48).
  • Documentary vs. DNA Genealogy: There are two main genealogical approaches—documentary, focused on written records, and DNA, centered on genetics. Charleen specializes in documentary genealogy, often collaborating with DNA genealogists for forensic cases (04:06, 05:04).
  • Practical Applications of AI in Workflow: AI assists Charleen in transcribing historical documents, filling out complex membership forms for lineage societies, and organizing daily tasks, which streamlines her workflow and increases productivity (06:02, 15:12, 19:09).
  • Human Oversight Remains Essential: Despite AI’s advancements, Charleen insists on triple-checking AI’s work, especially because errors—like confusing similar names across generations—remain common. Ultimately, her reputation depends on the accuracy of her work (26:10, 26:54, 33:44).
  • Ethical and Accuracy Concerns with AI: AI can amplify errors from historical lies or misinformation found in family stories and documents, making it crucial for humans to scrutinize both digital and written evidence for veracity (30:35, 32:19).
  • AI Does Not Replace Empathy or Context: Delivering sensitive family findings and dealing with emotionally charged or complex history requires human empathy and judgment, which AI currently cannot provide (34:10, 34:46).
  • AI’s Impact on Business Models: AI’s efficiency can threaten hourly billing models by reducing the amount of billable drudge work, prompting professionals like Charleen to rethink service structure, and workload management (16:39, 17:48).
  • Importance of Human Expertise and Tradition: The field still relies heavily on established human experts and the tradition of passing down nuanced industry knowledge, a depth of experience not yet replicable by AI (38:12, 39:04).
  • AI as a Creative and Motivational Tool: For Charleen, AI also serves as a way to break writer’s block and provide a starting point for her own edits, making the creative process less daunting (06:50, 07:04).
  • Caution and Balance in AI Adoption: Both Bob and Charleen agree on the need for a cautious, balanced approach to AI—utilizing its benefits without relinquishing critical control, ownership, or human touch in their work (40:17, 42:30).

Questions This Episode Answers

Q: How is AI transforming the field of genealogy research?

A: AI is making genealogy research much faster and more manageable by automating tasks like transcribing old documents, sifting through massive, previously unindexed archives, and filling out complex lineage society applications. According to Charleen Mullenweg at 06:17 and 15:01, this lets genealogists focus more on high-level research and less on tedious manual work, significantly reducing the time required for large projects.

Q: What are the main ways Charleen is currently using AI in her work?

A: Charleen Mullenweg shared that she uses AI for three main tasks: quickly processing large amounts of data, automating form completion for society applications, and organizing her workflow (18:13). While AI assists with these, she always double-checks its work due to possible errors or inconsistencies.

Q: Can AI completely replace professional genealogists?

A: No, AI cannot fully replace professional genealogists. As Charleen Mullenweg points out (24:37), human expertise is necessary for interpreting ambiguous records, understanding historical context, and ensuring accuracy, especially when dealing with complex family histories or conflicting data.

Q: What are some limitations or pitfalls of using AI in genealogy?

A: Charleen Mullenweg highlights that AI can mix up names, especially when dealing with multiple generations sharing the same name (27:04). It also cannot always discern fabrications or lies found in historical or family records, and there’s a risk of compounding errors or creating convincing forgeries (30:35), so human oversight is critical.

Q: How does AI handle old and handwritten genealogy documents?

A: Modern AI is increasingly capable of transcribing handwritten records from the 17th, 18th, and 19th centuries, which has greatly accelerated the document review process (11:42). However, AI’s accuracy depends on the quality of handwriting and the clarity of scanned images, so manual review is still required.

Q: Why is human oversight still necessary when using AI for genealogy?

A: Human oversight is essential because AI can misinterpret context, miss subtleties, and propagate existing human errors or fabrications in historical records (32:19). Genealogists like Charleen Mullenweg stress the importance of checking AI’s results and ensuring the information aligns with documentation and historical facts.

Q: How has AI changed the process of applying to lineage societies like the Daughters of the American Revolution?

A: AI can now assist by auto-filling lengthy applications using custom-trained wikis and instructions provided by genealogists (15:52). This automates much of the paperwork, allowing genealogists to spend more time on actual historical research and less on repetitive administrative tasks.

Q: Are there concerns about AI-generated forgeries or misinformation in genealogy?

A: Yes, Charleen Mullenweg expresses concern that it’s easier than ever to create fake or misleading documents using AI (30:35). As a result, genealogists must rigorously verify sources for authenticity and consistency within historical context to prevent the spread of misinformation.

Mentioned Links and Resources

  • Charleen.mullenweg.com (Personal Website) – Charleen Mullenweg suggests visiting her website to learn more about her genealogy and research services.
    🔗 https://charleen.mullenweg.com/
  • Courtney Robertson’s Blog (AI Wiki Application in Genealogy) – Charleen Mullenweg credits Courtney Robertson for a blog post detailing how to use an AI-fed wiki to instruct agents for research purposes.
    🔗 https://courtneyr.dev/2026/04/15/llm-wiki-ai-second-brain-obsidian/
  • FamilySearch.org (Free Genealogy Research Platform) – Charleen Mullenweg describes how FamilySearch, run by the LDS Church, provides millions of digitized records and utilizes AI for document searching and transcription.
    🔗 https://www.familysearch.org/
  • ProGen Study Groups (Professional Genealogy Training) – Charleen Mullenweg mentions participating in the ProGen Class 52, a program for professional genealogists.
    🔗 https://progenstudygroups.com/
  • Daughters of the American Revolution Genealogy Education (Lineage Society Training) – Charleen Mullenweg completed the DAR’s genealogy modules, which focus on lineage society documentation and research best practices.
    🔗 https://www.dar.org/national-society/genealogy

Timestamped Overview

  • 00:00 Specialization in Texas and New England history
  • 06:24 Using AI for initial transcriptions
  • 10:05 Challenges of Unsearchable Data
  • 11:12 Searching Historical Documents with Full-Text Search
  • 17:21 Adapting to industry changes
  • 18:52 Using document software for tasks
  • 23:32 AI’s impact on genealogy work
  • 27:42 Challenges of data disambiguation
  • 30:35 Using the Genealogical Proof Standard
  • 35:22 Using genetics for lineage verification
  • 37:53 Genealogy research touchstones
  • 41:45 Discussion on AI’s impact
  • 43:36 Using AI for design help
Episode Transcript

Bob Dunn: Hello, Bob here, and we are continuing AI for the Rest of Us. Initially, why I call it that is because all the tech people use AI in specific ways, and I thought, wouldn’t it be interesting to hear how it’s affecting other industries and different, yeah, just different places, and how people are dealing with it and what they see happening.

I have a great guest, and we’ll talk about what she’s going to be talking about soon, but I want to welcome Charlene Mullenweg.

Charlene Mullenweg: Hi, how are you doing?

Bob Dunn: I’m doing good. And yourself?

Charlene Mullenweg: Excellent. It’s hot, hot, hot in Texas right now.

Bob Dunn: Oh, okay. It’s not really hot here, but we’re going to have heat in Portugal this weekend, a little bit of a heat wave. And with the humidity, it becomes, you know, well, you know how that goes.

Charlene Mullenweg: Yeah. Brace yourself. It’ll be like coming to visit me in my hometown of Houston. But, uh, yeah, no, fortunately it’s not quite as humid here.

Bob Dunn: Yeah. Very, very cool. So the topic we’re going to be talking about is AI and genealogy. I want to have Charlene tell us a little bit about herself and what she does. She’s very immersed into this, and let me tell you, I can swear by this.

Charlene Mullenweg: You’re understatement of the year there, Bob. Uh, yeah, I’ve been a genealogist since the ’90s. I feel really old right now. And a professional genealogist for about 8 years. I hung out my shingle at that point.

I have a certificate in genealogical research from Boston University, which was pretty much the only degree I could get in genealogy without going to Utah or Scotland. So I chose the nice in-between. I’ve also done, um, there’s a series called ProGen, which is Professional Genealogists. I was part of ProGen Class 52.

And then for the Daughters of the American Revolution, they have an excellent series of training, mostly how to fill out their lineage society papers, but it’s an excellent genealogy training series as well. I’ve done all of their modules. So I’m pretty immersed in this.

Bob Dunn: You are.

Charlene Mullenweg: I would say I’m up to my eyeballs.

Bob Dunn: So before we get into what we’re going to talk about, what do you offer, or what do you provide? I don’t even know all the different facets, so I’m sure there’s some specific things you specialize in.

Charlene Mullenweg: I personally am specialized in colonial Texas history, to about the end of the Civil War, and then colonial New England history. Those are the 2 areas that I seem to have gone off into. Speaking Spanish and a little bit of German, enough German to be dangerous, is really helpful in Texas. But yeah, so I tend to deal a lot in colonial history.

Okay, so genealogy, you’ve got 2 kinds, I’m oversimplifying here, but you’ve got 2 kinds of genealogists. You have DNA genealogists, and they do cross over, but DNA genealogists, all they do is genetic work. It’s fascinating and it’s a little more complicated than I am.

And then you have documentary genealogists, people that work with the documents that are on hand. Now, obviously, you could say your great-granddaddy is so-and-so and the documents could back you up, but the DNA doesn’t. Sometimes what they call an NPE, or a non-parental event. It’s a very exciting thing, but it just means that somebody fibbed about who was the daddy or mommy in some cases, and it went sideways.

But my realm is entirely the documentary realm. I have worked with genetic genealogists several times to do, especially, forensic genealogy, where they’ll process an older DNA test, like from a crime lab or something, and they identify like a common 5th great-grandparent. You have to work forward and try and find who the victim or the perpetrator of the crime is. And I’ve done the work-forward part.

Bob Dunn: Wow.

Charlene Mullenweg: Yeah, ’cause that’s dealing with public record documents, birth certificates, death certificates, marriage certificates, court cases, that sort of thing. I live in the documentary world, and I spend a lot of my time in the 17th and early 18th century, so from like 1609 to about 1830. I tend to live in there.

And not to jump too heavily into AI, but there’s been a lot of really remarkable jumps forward in being able to do that work with AI. I’m no longer having to transcribe all by myself, you know?

Charlene Mullenweg: I’ll have AI transcribe the document, and then I’ll go over it with a fine-tooth comb, but it takes far less time than me doing a bulk transcription of a 17th century probate case and then having to fine-tooth comb myself. I’m more likely to miss things when I’m reviewing my own work than when I’m reviewing AI’s work.

It’s sort of like I get writer’s block if I have a blank page in front of me, but if I get AI to write something, it’s so much easier once I’m hate-editing, is what I call it. Once I’m hate-editing, like, ah, you wrote crap, I’m going to rewrite this whole thing, and I break that writer’s block.

Bob Dunn: Yeah, yeah. And that makes sense. That’s how I’ve used it a lot. But, you know, it’s interesting, little things have popped up here and there, and I’m thinking, you know, history documentation, the things you’re using it for, and you said how it’s advanced and helped, there seems like there’s other impacts that AI could have with genealogy that could really be significant.

Charlene Mullenweg: Well, look, I’m an old-school sci-fi-reading nerd. I grew up on Asimov and Clarke and Heinlein, Bradbury, and all these people. If you read them, you are expecting HAL 9000, you’re expecting Skynet, right?

And I avoided AI for a long time. I was just terrified of it, because if I allowed it any entry into my life or my work, it was going to take over and kill me, and then they’d find my desiccated body years later, and nobody would know, because I had been putting out output in my name.

But that is not the case. I found when I started using it that I had a very enthusiastic, kind of pedantic, and completely incompetent research assistant. It’s very helpful to give it a slog, just a pile of data that I would normally have to go through piece by piece over hours, days, and weeks of research. Just give it a slog of data, let it do a first pass.

I first encountered it at familysearch.org, which is run by the LDS Church, but is a really useful free research area for genealogists around the world. They first started using it to do something called full-text search. So what FamilySearch does is they go out and they’ll go to, I don’t know, Pawtucket, Pennsylvania, and they’ll go into their little clerk’s office and just scan all the documents up to a certain point, usually about 50 to 75 years before the date that they’re scanning, they stop, just for privacy’s sake.

But they’ll just come back with thousands of images, and they’ve been doing this since the ’60s or before. I mean, there’s some that are photographs from before they had scanning capability. It’s just genius what they did. And they were relying on volunteers to then go through this giant slog of data and index it so that it becomes searchable and usable by Charlene, you know, little Charlene sitting at her computer in the ’90s in Houston.

So they were entirely unindexed, unsearchable, everything. And if you wanted to use that collection at all before AI got into the picture, you had to literally go through, okay, I think they’re in Pawtucket at this point in time, so I’m going to pull this box of images. Oh, there’s only 900 that I have to go through. And you would have to go through them individually until you found the record that you were looking for. And you didn’t know what was out there other than a sort of vague descriptor that somebody may have put on the box of images in 1983, right? And it says basically Pawtucket records, and you don’t know what kind of records they are.

They could be birth records, they could be funeral home records, you don’t know. So they brought in what they call full-text search a few years ago, and what it does is it’s better at slogging through these millions, and I’m not even joking, multi-million unindexed documents, and sort of looking for names.

Now, of course, it gets better every day at reading 17th, 18th, 19th century handwriting 90% of the time, but it goes through all these and it starts picking out the names that you need. So, of course, when it started, you could only search for, let’s say, the last name Brooks, and then you could search for Brooks in Virginia, and then it was Brooks in Virginia in this county on these ranges of dates. So you started being able to drill down this giant search, and you’d start out, when I first ran Brooks, which is my great-great-grandmother’s name, in Virginia, I started out with over 120,000 documents.

Bob Dunn: Wow.

Charlene Mullenweg: For Sarah Brooks. And then once we started being able to drill it down, I got down to 700 documents, which sounds like a lot, but instead it’s a month of work instead of 5 years of work. Right.

So some of the available, but pretty much inaccessible, if that makes sense, data started becoming available. And then lately, I got really nerdy. So Courtney Robertson wrote this great piece on her, I think it’s on her blog, on putting together an AI wiki and having it inform the agent to do what you want.

I’m looking at it and I’m like, I could use that. A lot of what my clients hire me to do, like 85% of my work, is somebody will call me and they’ll say, I want to get into the Daughters of the American Revolution, or I want to get into the Order of the First Families of Connecticut, or something like that. And they may hand me another group that they joined on the same line, but it may be, you know, 5, 6 generations short of the goal that you need.

So I have to then learn the requirements for entry for the society that they want to join. I have to learn the sources that they find acceptable. Some places, if you say this, you have to prove it with a piece of paper. Some places they’re like, well, if you kind of logically get to that proof, then we’ll accept it. Some places are more particular than others about what documentation I can use, all this.

So I got a little inspired by Courtney’s work, because I called her and I said, Courtney, I could use this for genealogy. And she’s like, go girl. So I got inspired with Courtney’s work last weekend, and I built an AI wiki, where I fed the rules for a certain society into that wiki. I fed the blank forms, like your application for membership, and I started teaching it the particulars of how to fill out an application. And it’s funny, it called me absurdly fussy at one point.

Bob Dunn: I can’t imagine. That’s terrible, you know.

Charlene Mullenweg: I was like, look, you can’t be shrinking down the data. People are reading this on a piece of paper. Well, that’s absurdly fussy. Why aren’t they reading it online? Then you can make the font as big or as little as you want. And I’m like, you can’t be doing that. It’s a piece of paper that is received. So, you know, you have to really, and sometimes it mixes up things and sometimes it leaves off things, but I actually managed to fill out a 13-generations-long lineage society application.

With AI. I taught it the instructions, I gave it a reference document, and it filled it out. Now, is it exact and right? No, we’re still fine-tuning the process.

But I was like, this is insane. Instead of spending hours and hours filling out these forms, I could instead concentrate on doing the research while the machine is filling out the forms. My head just exploded.

In one way, it’s fantastic, because it means I can take on more projects and more clients. And another way, I get paid by the hour, so it’s good for my clients, not always good for me.

Bob Dunn: Yeah, well, that’s true, you know, and everybody seems to want to do those or have it do the menial tasks. I don’t want to say menial, but, you know, those kinds of things.

Charlene Mullenweg: The drudge work.

Bob Dunn: Yeah, the drudge work. And I find it interesting that it doesn’t click in their brains when they’re so into, you know, I’m gonna automate everything in my life or in my business, that there are people that do get paid by the hour, and that’s how they actually made a decent living, because, you know, they were paying you to do all that. And when that kind of gets cut out, that’s something a lot of people don’t realize is another, you know, whatever you want to call it, little side effect that can make you have to rethink, you know, do I charge more per hour? How do I make up for that? Do I need to make up for that because I can get more work? Or—

Charlene Mullenweg: Yeah, I thought it was going to really kind of quash my business model. And then I realized, right now at least, I’ve got a pile of work sitting behind me that I can now get to in kind of a reasonable time. But, you know, there’s trade-offs.

So, 3 main things, nobody expects the Spanish Inquisition, that I’ve used it for recently are slogging through giant piles of data. And of course, a lot of websites are limiting AI’s access to their content, which I’m 100% behind. If you don’t want AI messing with your stuff, you know, it’s your internet, you build it the way you want it to behave.

Bob Dunn: Yeah, exactly.

Charlene Mullenweg: Yeah. So, you know, I’ll get a pile of documents for a common last name like Duke or something, and I’ll just let it slog through that pile and slog through the reference documents until I find the line that I need. And then, of course, I have to double-check it.

The second thing I’ve used it for now, apparently, is filling out paperwork, the modern-day stuff, where I can spend more time finding new sources and new documents that say things. And then the third thing I’ve used it for is to kind of organize my life. You know, what am I doing today? Where do I have to go? I live in Post-it notes, so it’s certainly cut down my Post-it note consumption a great deal, which everybody’s very happy with because they’re everywhere.

Bob Dunn: They’re everywhere.

Charlene Mullenweg: I had to pick up a pile to do this podcast, but like, here’s some. Yeah.

Bob Dunn: And then you find one on the bottom of your shoe stuck to it and you think, oh, I was supposed to do that yesterday.

Charlene Mullenweg: Right. But so it’s really helpful for me on those 3 things. And only, well, all 3 of them touch genealogy, because my entire life and business is genealogy right now, but I can do more with it. I do have to check it, and I try not to get too angry with it, because it is a machine. It doesn’t care if I scream at it, but I care.

Bob Dunn: Sometimes you just gotta do it. Yeah.

Charlene Mullenweg: But yeah, it’s really been very helpful. It really has. Especially in tasks that would have taken me hours, like indexing a large spreadsheet of data or something like that. So we’ll see. My relationship with it has certainly become more warm where it was glacial before.

Bob Dunn: Yeah, that’s how I am. I’ve found the things that seem to work for me. There’s some very menial tasks that I still like to do or just feel like, you know, I like that time, because I don’t want my brain always rushing 100 miles per hour. I like to sometimes do a little work that just, you know, my brain’s not having to work a lot, but it does take a little bit of time.

With the fear that, you know, everybody’s having, what is the feeling within your community of genealogists, are is everyone gonna just start sitting down with Claude and say, let’s build out this, let’s you and me talk about it, and you’re gonna go and scan everything in the world and you’re gonna create what I want to know? Or is there too many working parts that maybe people don’t realize, and they think they’re going to get what they want, but they really can’t do it because they don’t know the actual process?

If you see what I mean, are any of them thinking, oh, AI is going to take over my job, or people will be able to do this sitting at home with a beer or a cup of coffee, or whatever their beverage choice is, and work with AI and get the information themselves.

Charlene Mullenweg: Well, I did have a relative of mine tell me at one point that AI did every single job that I’ve ever done in my lifetime, that AI could do it. And I was like, oh yeah, I was a bartender for a while. And he’s like, well, we could get a robot and put an AI on it. And I’m like, okay, whatever, nobody wants a robot bartender other than as a novelty.

Yeah, you know, it’s like the robot baristas, it’s all fine and good, but eventually you want a human touching your drink, you wanna speak to someone human. So I get what you’re saying about wanting to do the fiddly things. I am so OCD sometimes that to relax when I’m anxious, I alphabetize. I calm myself by alphabetizing something. I’m not even joking. I have a library, I’ve reorganized that library at least 5 times now, and I’ve lived here 6 years.

Bob Dunn: It’s almost a yearly thing, right?

Charlene Mullenweg: It’s almost a yearly thing, and it takes a few months to get through. I only allow myself like a bookcase at a time if I’m stressed. So I get that. There’s a lot of things that I still want to do, and there’s a lot of things that I don’t trust AI 100% to do.

When I hand off a product of any kind that I do to a client, my name is on that, and I had better check it, and I had better test the whole thing from soup to nuts if my name is on it. Because once your name is trash, your name is trash, you can’t fix that. I’m very wrapped up in that. So even if I managed to somehow make Claude or Gemini or ChatGPT do all my work for me, I’d still triple-check it. It mixes up names a lot.

Bob Dunn: I bet it does. Yeah. I’ve had some issues with that, and I can’t even imagine with what you’re working with, at the volume and the volume of names.

Charlene Mullenweg: Say you’ve got, I had a friend who’s passed away now, bless her heart, I miss her so much, she had 10 or 11 generations of men named Leonard Ballou. It was Leonard Ballou I, Leonard Ballou II. My father was Louis Charles Mullenweg IV.

I mean, when you’re dealing with 10 generations, or even 3 generations, of people that are named the same thing, and they’re all in the same area, you have to have a human go through there and really decide, okay, this occurred on this date, so it could only be one of these 3 Leonard Ballous. And, all right, now it says the wife is this one, so it could only be this second Leonard Ballou.

You really have to, the fiddly things, the little tiny nitpicky things, it’s gonna get wrong every time. And if I were to write that specific ability into it, what would I then be losing? If I write an instruction too explicitly, could I then be losing data that could be useful because it doesn’t fall within these strict parameters?

Now, on the other hand, it notices things that I’m skimming through, dozens of documents, and I have been trained to skim through dozens of documents really, really fast, even handwritten 17th century god-awful writing. I’m trained to go through it very fast and pick out words that are important to what I’m researching, my research question, we call it. And it picks out things that I don’t see, because I’m looking for a specific set of parameters and it’s just reading the document. That’s helpful, that’s extremely helpful.

And my other worry is this: people, when they’re talking about their family trees, somebody’s often told a lie. Now, that lie could be, you know, oh, so-and-so, Max is David’s daddy, and it could be the mom was telling that lie to protect herself, because adultery was punishable by banishment or branding, you know, we’ve all read The Scarlet Letter. There’s 100,000 reasons for people to lie, and they will often do so, especially as you go further back in history, not always that far back.

They will often do so in documentation. I had a great-grandmother that never told the truth about her age, on any single document, her entire life. I must have 30 documents for her, and every single one, her age is different. That being said, it’s supremely easy, even before AI, to make a fake document. It’s horrible to say, I get sick thinking about it.

Bob Dunn: Yeah.

Charlene Mullenweg: But people have done it for centuries, and with AI, making a deepfake is extremely easy and very frightening. And one of our jobs is to judge the veracity, the reliability, of a source, and make sure that it fits into what we call the GPS, the Genealogical Proof Standard. You all get global positioning satellites, we get Genealogical Proof Standard for GPS.

So I have to look at a document and make sure that it’s correct with the time period, there’s no anachronisms in it, that it’s correct to the family that I’m studying, that it’s correct to the location that I’m working in, as best as I can. Is it a true document? Is the paper correct? Is the ink correct? Sometimes it gets that particular. Does the handwriting match the clerk who wrote a document the day before and a document the day after? I could train AI to do that. I don’t want to. I, again, have to put my name on it when I hand it off, so I would be terrified to. Could I have it make a pass at it? Sure, no problem.

Bob Dunn: And, you know, basically the lies and those fabrications humans make could compound AI hallucinations. I mean, they can—

Charlene Mullenweg: They really do.

Bob Dunn: They could take that one lie and blend it into other things, because to them it’s not a lie.

Charlene Mullenweg: And family stories, whether they’re true or not, can often be published, especially around the turn of the 20th century. There was a lot of what they called mug books. They would print these mug books about prominent citizens in the area, and they would tell the stories of their family, how they got there, you know, oh, John Paul Getty got here, his family got here on this date, and they did all these great things before it got to him, whatever, right?

And people would often pay to be in these books, like Who’s Who, and all that. And therefore you have a published source telling you one fact, and you have actual documents on the research ground telling you a different fact. How do you weigh those against each other?

So I often say, look, a family story is a fairy tale until I get some documentation around it, right? Drives my poor mother crazy. She’s like, no, I’ve been told this all my life. And I’m like, but actually, the documents say that it can’t be so. She’s like, uh-huh, but that’s a distinction that I think a human still needs to make, and I would be uncomfortable entrusting that sort of delicate work to a machine.

I had a client who was told her grandfather died in the late ’20s, and it turned out he lived until the 1960s.

Bob Dunn: Yeah.

Charlene Mullenweg: And I had to go back and explain this to her, and her father, and I just accidentally uncovered it. But, you know, I wouldn’t have an AI write that email. You have to have some care and some empathy and some understanding of the fear that these people were maybe experiencing, or the danger they may have been in.

Or even for something that sounds trivial to us today, like, oh, so-and-so may have gotten shunned from society if she didn’t tell this lie. Well, that sounds trivial to us because we can always find another society, but her society, it’s her world, it’s everything. They didn’t have internet, they didn’t have ways to find people like we do today. So you have to be very careful in dealing with these long-entrenched and sometimes publicly published family stories and how you present the actuality to people, or as far as you could tell from documents. Now, genetics may tell an entirely third story.

Bob Dunn: Let’s throw that in, right?

Charlene Mullenweg: You know, one of the lineage societies, one of the big ones, started using genetics. Everybody used paternity tests since they were available. If you had a court order that said, you know, there was a paternity test done that says your daddy was your daddy, well, then we’re good, right? And they’ve used that.

One of the larger groups started using genetics, like DNA, incredibly well. And I’m just like, oh God, you know, you say your ancestor’s Benjamin Franklin, and then you find out, no, no, it was, you’re not a Franklin at all, or you’re not a Washington at all, or you’re not a Jefferson at all. I mean, the dirt that it uncovers. You just never know what’s going to happen when you actually put genetics into the picture. It’s just like, oh God, how did that happen?

Bob Dunn: Yeah, you’re not dealing with just a bunch of facts, you’re dealing with human life. It’s kind of mind-boggling to me, all the different details and how you pull those together from one side to the other, how you disprove something, even though you’re not saying, I want to disprove it, it’s just, we gotta know if it’s right or wrong.

Charlene Mullenweg: Yeah, and sometimes lack of evidence is not evidence of a lack of something. It’s like, sometimes the things you don’t find are equally important as the things that you do find. Like, you don’t find a marriage record, why is this? How important is this?

Your listeners can’t see, but I have bookcases and bookcases full of books behind me, and I use at least 3 of them a day. Everything that I don’t touch, I pretty much have it electronically and I’ve sold the paper copy. But I live in these books, and what it is, is compilations of who came when, what we actually know about the family, things like that.

And so, like, when I was talking about sci-fi earlier and I’m dropping the names Asimov and Heinlein, you know who I was talking about.

Bob Dunn: Yeah.

Charlene Mullenweg: These are touchstones. Well, in genealogy world, we have Anderson, we have Roberts, we have Jacobus, oh my God, Jacobus is huge for Connecticut, at least. We have these names that we know we can return to when we need information, because they’re the experts in our field.

The Richardsons and the Roberts and the Jacobuses, they are our touchstone when we can’t find data. And we know, inherently, we learn as you do this for decades, kind of where their strengths and weaknesses are. Like, Jacobus did research all over New England, don’t get me wrong, but I trust him most in Connecticut. Why? Because that’s where I feel he did all his best work. You learn their strengths and weaknesses as you go along.

I just don’t have the time or the patience to sit there and teach that item by item to an AI. My fear is the next generation loses that knowledge because nobody sat there and taught it to an AI.

Bob Dunn: Yeah. Wow. That’s amazing. You know, we’re still gotta have that human touch, that we’re still trusting people, even if that person is putting together a bunch of stuff using AI. We’re trusting that they’re still giving us what we expect, or what we paid for, or what they promised, and they need to make sure it pans out that way.

Charlene Mullenweg: Yeah. And I would say I’m confident in the AI stuff that I’ve managed to do, but I don’t know that I’d hand much more over to it. When I give a presentation, it has to be my words.

Or my reputation just goes out the window, right? I might have AI help me edit my proposal or get me started with writing prompts, but I think at some point we need to own it, no matter how much help we have on getting there. I don’t feel comfortable owning something until I’ve touched it, I’ve handled it, I’ve reviewed it, I’ve looked into it. So it’s a cautious approach.

Bob Dunn: Yeah, I love it though. It sounds very similar to my own way of thinking around AI and—

Charlene Mullenweg: Well, I love that you’ve asked me about it, because so many people are like, oh, you’re a genealogist, huh, that sounds interesting, but they’re lying. It doesn’t sound interesting to anybody at all.

Bob Dunn: Well, it is. It’s incredibly fascinating, and it has me thinking about all different kinds of things, and, you know, it’s a lot of work you have to put into it. It’s not just reading and writing.

Charlene Mullenweg: I have an uncle who calls me periodically and he says, I’m having trouble falling asleep, please tell me about the family tree, that’ll cure me right away. And then he laughs maniacally and hangs up. I mean, I could put everybody to sleep with this, and I’m trying not to.

Bob Dunn: Oh yeah, yeah. It’s why I started this series, it brings out what I think is the interesting part of whatever profession, because not only do you learn about it, but I think there’s a lot of ways, since everybody’s into AI, well, not everybody in the world, but a lot of people are into AI, they can really relate to certain parts of that, because somehow that reflects on maybe something they’re doing or something they’re thinking of doing with AI. And maybe rethinking it, thinking, you know, am I, I gotta maybe pull back on some of it, I may need to look at different parts and make sure, like you said, own it.

Charlene Mullenweg: You gotta own it. Well, I was surface level before, and now I’m 20,000 leagues under this, and I’m a little confused at what I’ve done, but kind of cool too.

Bob Dunn: Yeah, exactly. Yep. Excellent. So now, what if somebody wants to reach out to you, or just learn more about you or the services you offer, where do they go to do that?

Charlene Mullenweg: I would send them to the website I’ve kept up with recently. Look, web design is not my game, but there is one, I call it the Tale of Two Charlenes. There’s the business-analysis-informatics Charlene side and the genealogy-and-research Charlene side. But I recently built out charleen.mullenweg.com, and I know you’ll put a link for me.

Bob Dunn: Yeah, definitely.

Charlene Mullenweg: And I’ve really kind of had fun playing with that. AI helped me, I will say, it told me what to do, because I’d be like, I can’t figure out how to make this box move this way, and it said, oh, this is where you do it on WordPress, or this is where you do it. So it really helped me. Design is not my deal. I’ve lived in databases my entire career.

Bob Dunn: Well, this has absolutely been wonderful. I’ve really enjoyed this, and yeah, it’s given me some glimpse into more of what you do. And I’m sure it’s very, very fascinating work you’re doing.

Charlene Mullenweg: Well, it’s kind of the story of who we are, but it doesn’t define who we are. It’s just another facet of each person’s life, their family history. I think it really helps folks. And certainly if somebody wants to get into a lineage society in the United States, I’m happy to help.

Bob Dunn: Well, thank you, Charlene. Love it that we were able to do this, and I appreciate you taking the time.

Charlene Mullenweg: Thank you so much, Bob. I’ve just had a blast.