RLP 357: AI and FamilySearch Full-Text Search

Nicole (0s):
This is Research Like a Pro, episode 357 AI and Family Search Vol Text Search. Welcome to Research Like a Pro a Genealogy Podcast about taking your research to the next level, hosted by Nicole Dyer and Diana Elder accredited genealogy professional. Diana and Nicole are the mother-daughter team at FamilyLocket.com and the authors of Research Like a Pro A Genealogist Guide. With Robin Wirthlin they also co-authored the companion volume, Research Like a Pro with DNA. Join Diana and Nicole as they discuss how to stay organized, make progress in their research and solve difficult cases. Let’s go.

Nicole (41s):
Today’s episode is sponsored by newspapers.com. Hi everyone. Welcome to Research Like a Pro.

Diana (48s):
Hi Nicole. How are you doing today?

Nicole (51s):
Doing great. I’ve been making some cool discoveries on how I can use AI to help me with pedigree triangulation for DNA work.

Diana (59s):
And that is super exciting because that’s so time consuming. So I am excited to try the same thing and to learn from you about what you’ve discovered.

Nicole (1m 9s):
Yeah, I was just trying to like get everyone from one cluster into a spreadsheet with DNA GEDCom and that actually worked great just using the Ancestry groups and then using DNA GEDCom to kinda gather all of their trees into one spreadsheet and then you could analyze the spreadsheet on your own or you could delete out the match names and put in some fake names and then upload it to ChatGPT or something and then have it analyzed looking for duplicate ancestors like who could be a common ancestor.

Diana (1m 38s):
I really love that. I am just thinking of all the ways that we can use AI in our DNA work. Because of my project that I’m working on right now, I am finally to the research stage, and I’ve been working on documentary work and my next thing to do is really diving into the DNA and I have this big Gephi graph with several clusters that I really want to explore and this sounds so much better than doing it manually where I just go click on a match. You know, I use a spreadsheet and find the matches that have trees and then go look in their trees to try to find common ancestral couples. So it sounds like a perfect use for AI to help us do that, so I’m excited to try it. Thank you for sharing a little bit about how you’ve been working on that.

Nicole (2m 21s):
Yeah, I’m just hoping to streamline the process and make it a little bit faster. It’s always something that I’ve wanted to figure out a way that was like, efficient. Like you said, clicking on each match and opening their trees kind of feels slow, but it does require that you think a little bit about privacy and things because unless you’re using like a temporary chat, the names of the matches would go back to the people at ChatGPT who are you reviewing things. So for privacy’s sake, I just deleted out the names of the matches and just left in the match ID number instead. That’s how I did that.

Diana (2m 56s):
That’s a good idea.

Nicole (2m 57s):
But yeah, what else are you working on?

Diana (2m 59s):
Well, I just started, like I said, my documentary research yesterday and number one on my research plan for this William Weatherford in Arkansas was to use FamilySearch’s Full Text Search. And I discovered in my locality research that the parent county of Izard County where my Henderson Weatherford was located, the parent county of Izard is Independence County. And so I decided that perhaps I should research independence for William, who’s the hypothesized father. And I thought, why not just put it into Full Text Search? Because we always wanted to start with the most likely to find an answer and the easiest, something we can do from home, and of course I can just do that from home.

Diana (3m 47s):
And I had seen that FamilySearch did have a lot of records from Independence County, and right away I got four hits for William Weatherford and clicked into those, they’re court records in the 1820s. And so he was building a road, he was on a jury a couple times and these seemed to be like the perfect time, you know, when it was, this area was Independence. And then he shows up in a tax record in Izard County in 1829 and the county was created about that time. And so I am thinking that probably he may have not even moved.

Diana (4m 31s):
They just had a new county created and then his records moved into that county or would be in that county. And one of the tricky things about Izard County is that it is a burned county. And so I only have federal land records and the state tax records and the federal censuses. So I have no deeds, no probate, no court records, nothing on the county level. So finding William in a county that actually has records was really exciting because I don’t have that in Izard County.

Nicole (5m 1s):
Yay. What a good idea to look at the parent county. Your locality research really paid off.

Diana (5m 6s):
Yeah, it did. And you might say, well, why didn’t you do that before? Well, I was busy on other things before, you know, this phase of research finally got to researching William. So in my last phase, I still thought that Henderson was part of the Dallas County WeatherFords, and then when I disproved that, then I could move to William and and learning more about him. So you know, it’s just phase by phase gradually getting closer to the answer.

Nicole (5m 36s):
Remind me where you found this William Weatherford as a hypothesis for Henderson’s father.

Diana (5m 42s):
He is on the 1830 census in Izard County, and he does have these three young males in his household, and one could be Henderson and then the two others look like brothers. And they appeared, they all appear in the tax records in the 1830s. And so William was old enough, you know, to be the father of Henderson. And then he’s in the area in the same time period there in the 1830s. And so he is the only Weatherford in the area that could be a father for Henderson. So you know, it’s just the most likely event.

Diana (6m 22s):
It could be that Henderson just showed up there without his family, but I need to explore this possibility that this is his family, this William and these other two that could be his brothers. So that’s where DNA will be coming in to see if I can make some connections through DNA as well.

Nicole (6m 40s):
Well, that’s fantastic. I’m excited to read more about it in your research report.

Diana (6m 46s):
Yeah, it’s fun and it’s exciting to actually have a hit from Full Text Search that would’ve taken me quite a while to go through the court records. Those are usually not indexed unless you have someone who has written a book and gone through and abstracted all those out, which I do have some books like that. But in this case, Full Text Search was just amazing and the transcription was really good. It was spot on. That was good.

Nicole (7m 16s):
It’s really exciting to have these opportunities to find our relatives in hard to research record books that aren’t really indexed like court records, so yay. And we’re gonna talk more about that in today’s episode, which is about the Full Text Search.

Diana (7m 32s):
Yes, I’m excited to talk about it.

Nicole (7m 34s):
Well, before we do that, we have a listener spotlight. We got an email from our friend Nancy, who is a podcast listener. Nancy said, “your podcast about Anna Nash and the Cowderys episode 355 was really interesting. My ears always perk up when you talk about Kirtland, Ohio as I lived in the Cleveland area, Geauga and Lake counties growing up. So I know of Kirtland. I wanted to give you an assist in pronouncing the name of the county.” I pronounced it incorrectly in the last episode, and so she was helping me out. The way that I pronounced it was gee, uga, it’s spelled G-E-A-U-G-A. And I do remember we talked about not knowing how to say it, and she says, “it looks to be like a bigger tongue twister than it is.

Nicole (8m 18s):
It is pronounced jaw-gah or gee-aga, jaw-gah, ghee-aga, depending on how fast you say it.” G–aga, I’m not sure. I think I think it’s gonna be a Juh. So that was fun to get that. So I think I will say gee-aga from now on instead of Gee-Uga County Ohio.

Diana (8m 41s):
Well, it’s so tricky when you don’t live there, you don’t hear anybody say it and you’re just reading it. And we have places like that everywhere in every state here in Utah we have a Tooele County and it looks like it should be tulle. You know, who, who would know? It’s Too-willa. So it’s just local pronunciation is always fun to know,

Nicole (9m 3s):
Right? So if you ever hear us mispronounce a location name, definitely email us and we will share that correction on the next podcast episode. Alright, for announcements, we have our next Research Like a Pro webinar in our Case Study Series on May 17th from Jessica Taylor Morgan, Accredited Genealogist. And the title is Avoiding Assumptions: Tracing a Family in 19th Century London. So she’ll be talking about a research project focused on George White who was born in 1863 and died in 1928 in London. And this will explore the varying availability of historical records, the diverse occupations typical of the period, and the challenges presented by misleading records.

Nicole (9m 47s):
So if you’re not registered for our 2025 webinar series yet, we invite you to register for that and join us on May 17th to learn from Jessica. Also coming up, I am coordinating the Texas Institute of Genealogical Research course about AI called, Integrating AI into Genealogical Research and Writing. So the Texas Institute of Genealogical Research, otherwise known as Tigger, is a week long opportunity to dive deep into a topic. And there are some other courses available like legal research in Texas and Southern research. And so if you want to join me and a team of great instructors, we will be talking about using artificial intelligence in our research and writing.

Nicole (10m 32s):
Registration is ongoing and it ends on June 6th. And the instructors include myself and my mother, Diana Elder, also Alice Childs, Mark Humphreys, PhD, Jan Joyce, Steve Little, Angela McGee, Katherine Schober, and Mark Thompson. And they’re all wonderful and they all have some great things to share about how we can use AI while still maintaining accuracy in our research. Also, we have the upcoming National Genealogical Society Conference in Louisville, Kentucky. Mom and I will be there on May 23rd through the 26th, and we will not be having an exhibit hall booth, but we will be doing a research trip on the first day enjoying all the classes and hopefully seeing several of you there.

Nicole (11m 18s):
So that’s coming up on May 23rd through the 26th. Also, I can’t believe it’s almost summer because that means that it’s almost time to start thinking about the next Research Like a Pro study group, which begins at the end of the summer on August 27th. And you can register for that beginning on May 21st at 10:00 AM and registration will conclude on August 21st. So be thinking if you’d like to join us, and we would love to have any repeat study group members come as a peer group leader and enjoy free registration. So if you’re interested, just email us.

Diana (11m 50s):
Well, it’s exciting to think about the summer and all the fun learning opportunities. I know I’m excited for the Institute class on AI and thinking back to last year is when we did our Research Like a Pro with AI workshop and learned so much and just really helped me to just get going with using AI. I’d been a little bit reticent before that and now it is just so much better than it was even then. So really fun. Well, and today we’re going to talk a little bit more about AI because we’re going to talk about FamilySearch’s Full Text Search, where they have used AI to create this amazing opportunity for us to find our ancestors.

Diana (12m 33s):
I think one of the challenges with our research is sometimes we do not know where our ancestor might have some records, we know the basic locations, but what if they have a record somewhere that we hadn’t thought to look? And that is one of the things that Full Text Search is so wonderful for. Well, it was introduced at RootsTech 2024, so it has been around for over a year now. And when it first started, I jumped on and started using it to discover more about my ancestor Thomas Beverly Royston. I didn’t hit anything new for him at that time. Everything that it showed me I already had, or I just didn’t find you know much for him.

Diana (13m 16s):
But during RootsTech 2025, I was playing around a little bit with it and just did it random search. And this time I found a new record and he was in a new location and it made a new family connection. It was really an exciting one. And Nicole, we were in our hotel room and I think we were both playing with Full Text Search and finding discoveries, and it was just kind of fun. So we are continually finding new record collections added to Full Tech Search. When it started out, you know, it was just land and probate and now they’re just opening it up to all sorts of different records.

Diana (13m 56s):
So we do need to remember to keep checking back because it’s not just static, it’s not just gonna stay. It’s going to keep growing with the collections they’ve added.

Nicole (14m 7s):
Absolutely. And they have grown the number of names indexed in just one year to be over so many million, I don’t even know the number, I just know that it’s increased exponentially in one year, which is so exciting. And so to find where to use Full Text Search, you’ll start at FamilySearch’s homepage. From there, scroll down to Family Search Labs and then click on View Experiments. Next, find the experiment labeled, Expand Your Search with Full text, and then click Go to experiment. The next screen allows you to start your search or search within a specific collection.

Nicole (14m 48s):
FamilySearch then provides some advanced search tips to narrow down the results and be sure when you’re doing this to experiment and play with it. Try different search terms, try removing some search terms, and then using the filters instead. And there’s a lot of different ways to go about it. So you can try a lot of things. So when mom put in Thomas B Royston in quotes and entered it into the keyword search instead of the name search, this resulted in 46 hits for several men named Thomas B Royston, and one of them was our ancestor. And sometimes this way of the keyword search in quotes works better than a name search. So you have to just try both and see what you can come up with.

Diana (15m 31s):
Right. And it was interesting when I looked at those results that a lot of them were from England, and I knew he was never in England, and some of them were well before or after his lifespan. And so I needed to narrow this down. So I specified United States and gave a year range of 1806 to 1868, his birth to death, and then it cut that more than in half and it resulted in just 22 results. Eight of those were actually for my ancestor, Thomas Beverly Royston. And when I started looking at those records, those eight records, only one of them was a record I already had.

Diana (16m 14s):
And so that was exciting. This was all new information. And so in the records in the counties where I knew Thomas resided, he was appearing as a witness or an estate appraiser. So he hadn’t been named in the main indexes of the probate or deed books that I had previously searched. And one record stood out and it was in a brand new location that I had not researched. That was Clark County, Georgia. And you know, we don’t know what we don’t know. I just had no reason to believe he would appear in a record there, but I was wrong. And it turns out that Clark County was the neighboring county to the counties he did live in, in Green County in Morgan County, Georgia.

Diana (16m 59s):
And you know, we, we often say, you know, look at the neighboring counties, but I had so many records on him already I hadn’t, I guess, been motivated to go, you know, seek out more records in the neighboring counties. So not only did Thomas appear in the record, but it also gave me some new information on his family. So this was really exciting. It was actually a mortgage, it was called Clark County Mortgages, and it was an indenture and it was David Irwin and Thomas B Royston. So we’ll talk a little bit more about what it specifically said later on in the podcast, but for a minute let’s talk a little bit about what it does.

Diana (17m 45s):
So as I mentioned, Full Text Search is AI powered. And so AI is reading these handwritten documents and looking for your search terms within them. When you click on the item, you see the image with your search terms highlighted on the page. And I love that because you don’t have to go through all the handwriting, try to find the search terms, you can just quickly see what it is and see if it applies. But then the other really neat thing about it is there’s an AI transcription to the right of the document. So you can actually read in text and quickly evaluate whether it’s for your ancestor or not. So I recognize several names in the document and knew right away this was for my Thomas.

Nicole (18m 31s):
It’s exciting when you can recognize the FAN club right away. So seeing those names adjacent to your target name, it’s exciting. Well, let’s have a word from our Sponsor. Today’s episode is sponsored by Newspapers.com. Break down genealogy brick walls with a subscription to the largest online newspaper archive. Did you know Newspapers.com has over 1 billion pages of digitized newspapers dating back to 1690? Their growing collection includes papers from the US, UK, Canada, Australia, and beyond. Discover birth and marriage announcements, obituaries and everyday stories about your ancestors in seconds. Newspapers.com can help you fill in the gaps between vital records and reveal details about your ancestors’ lives that you can’t find anywhere else.

Nicole (19m 13s):
Their easy to use search feature lets you filter your results by date, location, specific paper and more. When you find something interesting, Newspapers.com makes it a snap to share it with family and friends. You can even save it directly to your Ancestry tree. Come explore 1 billion pages and make infinite discoveries today on Newspapers.com. Use promo code FamilyLocket for a 20% discount on your subscription.

Diana (19m 37s):
Alright, well let’s talk next about transcribing the document. So sometimes the AI transcription in Full Text Search is really good and sometimes it’s not so great. It, I think a lot of it just depends on the handwriting. In this case it wasn’t so great. So I used additional AI to review the document and provide the transcription. So first I took a screenshot of the second half of the image, which was the applicable document because this was a two page spread. So I just screenshotted the part I wanted and then I gave it to both Claude and ChatGPT and both models struggled with the handwriting and preserving the line breaks.

Diana (20m 20s):
So I’m not sure why, they don’t always struggle, but in this case, they were struggling. So next I decided to try Transkribus because it reads each line and preserves the line breaks. So in the blog post that I wrote, I have an image showing the dotted line on the left that corresponds to lines on the right in Transkribus. And so it’s really nice because it does help you to check the transcription. And Transkribus actually did a pretty good job, but it needed editing. So the nice thing about Transkribus is you can edit the transcription right there, or you can copy and paste it elsewhere. So I decided this was a good opportunity to have AI check AI.

Diana (21m 1s):
And so I took Transkribus’s transcription and copy and pasted that into ChatGPT. And it already had the image and I asked it to preserve the same line breaks and correct the errors. And this result was very good. And so I could add any final edits myself because ChatGPT showed the transcription in Canvas and in Canvas you can edit it. So once I had a good transcription, the next step was to try to understand the document because this one was a little challenging, a little interesting. So that’s the thing that I love about AI is once you have got your document figured out, you know through transcription, then you can start learning about what it’s actually telling you.

Nicole (21m 46s):
Yes, that’s wonderful. Well, let’s talk about what was actually in this record that was found. Let’s read the summary and explanation of the indenture by ChatGPT. So ChatGPT said this indenture dated February 6th, 1830, in Clark County, Georgia is a marriage settlement agreement between David Irwin, Sarah B Royston, and Thomas B Royston. It is a legal contract concerning the ownership and future inheritance of enslaved individuals in anticipation of the marriage between David Irwin and Sarah B Royston. And then next ChatGPT breaks it down into some different sections. So it says, “Key provisions in the document.

Nicole (22m 26s):
Number one, parties involved David Irwin, first party, the prospective husband, Sarah B Royston, second party, the prospective wife who owns Property, Thomas B Royston third party, trustee responsible for ensuring the agreement is upheld. And then the second section is property in question. And then under that we have a couple bullets. Sarah B Royston possesses Isaac, a boy about 10 years old, and Thomas, a boy about six years old, in Fee Simple, which is full ownership, these individuals are identified as part of her personal estate.” Then number three, ChatGPT gives us a section called Purpose of the Agreement.

Nicole (23m 9s):
And then it says, “David Irwin does not have an estate sufficient to provide a financial settlement equivalent to Sarah’s property to ensure Sarah retains control of her property. A trust is established in favor of Sarah with Thomas B Royston as trustee. Then for terms of the agreement, the enslaved individuals will remain under the use and benefit of Sarah B Royston during her lifetime. After her death, they’ll pass to her children. If she has no surviving children, she retains the right to dispose of them as she wishes.” And finally, section five is legal, witnessing and recording. The document was witnessed by John Dawson, Robert Cessna, and Betsy Ann Dawson.

Nicole (23m 53s):
It was sworn before Justice of the Peace, S Brown, and recorded on August 5th, 1830. Then ChatGPT gives us a section called Historical and Legal Context. And first it says marriage and property law. At this time, married women often lost legal control of their property to their husbands. A marriage settlement like this was a way to protect a woman’s assets, slavery and inheritance. Enslaved individuals were legally treated as property and were often included in estate and marriage contracts. Highlighting the legal structure of slavery and the Antebellum South trustee role, Thomas B Royston was designated as the trustee, meaning he was responsible for ensuring that Sarah’s property rights were upheld.

Nicole (24m 37s):
And then ChatGPT’s conclusion was this document is a marriage contract and property trust ensuring that Sarah B Royston retains control over her property, enslaved individuals, and that they pass to her heirs rather than becoming David Irwin’s property upon marriage. This agreement reflects antebellum’s southern legal practices regarding women’s property rights inheritance and the institution of slavery.

Diana (25m 2s):
Right. So that was really nice because a lot of times, even in current legal records, it can be hard to understand what’s actually being said and what what it means and how it all works together. And so in this case, it was very, very helpful for ChatGPT to give me an idea of what was going on. So through my research originally on Thomas B Royston, I had concluded that Sarah was his sister and Robert was his half-brother. However, they were never mentioned in a document together. So with Thomas acting as her trustee, I now have another piece of evidence of the relationship and I was really interested to see John and Betsy Ann Dawson.

Diana (25m 46s):
So I would love to do a complete project on John Dawson and this Betsy, because I do have a Betsy mentioned as another half-sibling to Thomas. And I also have John Dawson who appears in all these other documents with the Roystons and Cessnas, and I’ve always wondered about his connection. So how is he connected to my Thomas B Royston? I don’t know. But it gives me another avenue for research and all because Full Text Search pulled up that record in a place I hadn’t thought to look and gave me another piece of the puzzle. So if there’s anyone listening who’s never tried Full Text Search, I would encourage you to try it out and see what you can discover because it truly is just a really wonderful way to do your research and find your people in unusual places.

Nicole (26m 40s):
I love it. It’s so exciting to get the name of Betsy Ann Dawson and just to find one more person to this family who’d researched for so long and and have that evidence of them together and just understanding the family that much more. It’s so neat to see. I just love seeing the women’s names written down when you don’t see them very often, And so it just is so nice.

Diana (27m 2s):
Exactly. Especially when it’s early, you know, when it’s just so early, pre-1850 when we don’t have them on censuses and we just get lucky if they’re in a deed or a witness or a probate record. So it is really great to get them listed.

Nicole (27m 19s):
Absolutely. And then even more difficult to uncover are the names and identities of the enslaved individuals who are mentioned. And it’s always exciting to find a name of an enslaved person. And that was another really exciting part of this, is getting those names and learning about them and considering ways to share that research and putting it out there, making it possible. Of course, putting it in this blog post helps.

Diana (27m 45s):
Yeah,

Nicole (27m 46s):
Whenever those names are mentioned, it’s good.

Diana (27m 48s):
It is. And these are the first names of enslaved people I’ve been able to discover from this family. I have looked at so many deeds, so many things, hoping that I could find perhaps a listing of all the names. But all I had so far is the slave schedules of 1850 and 1860, that list, you know, whether male or female in their ages, but no names. And emancipation happened before Thomas wrote his will. So there’s no names mentioned in any of his probate. And so I, I would love to be able to trace down those people and do some research. So this at least gives me two that are in this document.

Nicole (28m 27s):
Right.

Diana (28m 27s):
It’s good.

Nicole (28m 29s):
It is, yes. And we know that because of this marriage agreement that if Isaac, the 10-year-old, and Thomas, the 6-year-old, were still living when Sarah passed away, they would’ve been passed on to her children probably the most likely situation, unless they had hard times or financial situation they needed to sell them.

Diana (28m 49s):
Yeah. I will have to now go do some research on her husband David Irwin. He ends up in Cobb County, Georgia and you know, she doesn’t die again, until after emancipation. And so you know, they, these people, these, this Isaac and Thomas by this time would’ve been adult men and would’ve been freed. So yeah, whether we can find out what happened to them or not, I don’t know, it would just require more research, which sounds like a really good thing to do.

Nicole (29m 20s):
Yeah, it would be easier to track them if she didn’t sell them before emancipation and then you could look for where they were living during emancipation and then just looking for Irwin Surnamed individuals, unless they kept Royston as the surname, or maybe they took on a new surname, which is the one of the biggest challenges, right, of enslaved individual research. But it is good to consider all these possibilities in different ways to research the family more.

Diana (29m 47s):
Right. This one record holds a lot of clues and a lot of possibilities.

Nicole (29m 52s):
Well, we just have to thank Full Text Search at Family Search and it’s so exciting to see so many more records becoming available and if you haven’t tried Full Text Search in the last few months, you should try it again. If you’ve never tried it at all, you should definitely try it. It’s making so many records available to us and it’s not only in English, it’s also in many other languages. So Spanish, Portuguese, and I’m not even sure all the languages that are available except for, I know that they do have Spanish because I’ve been kind of playing around with transcribing Spanish language records from the 1600s from Ecuador with AI and it did amazing.

Nicole (30m 36s):
I was using ChatGPT, but then I was comparing those transcriptions to the ones done by Full Text Search and they were both really good, but they had some differences. And anyway, it’s been interesting to compare that and it’s, it’s exciting to have AI being able to do transcriptions for us. So Hooray,

Diana (30m 55s):
Thank you. And getting so much better. We still have to check things a lot, but it’s nice to see for the most part getting much better.

Nicole (31m 4s):
Yeah. Oh, I finally found the article that I was looking for. So earlier we mentioned that their Full Text Search has been increasing the number of records available. So there’s a blog post at the Family Search blog called, Full Text Search Find Records Faster with the Help of AI. And this came out last month on March 19th, and it’s talking about how there are now over 1 billion more historical records because of this handwriting recognition technology at Full Text Search. So 1 billion, it’s just such an enormous number. It’s hard to even fathom it.

Diana (31m 38s):
That’s amazing.

Nicole (31m 39s):
In just one year, this technology has made it possible to find that many more people or instances of these people. So we will put the link to this blog post from Family Search into the show notes as well so that you can learn more about it there and hopefully share it with all the friends and family history you have so everyone can make more discoveries.

Diana (32m 1s):
Yes, absolutely. Well have fun everyone. Thanks for listening. We will talk to you next time. Okay. Bye bye-Bye.

Nicole (32m 46s):
Thank you for listening. We hope that something you heard today will help you make progress in your research. If you want to learn more, purchase our books, Research Like a Pro and Research Like a Pro with DNA on Amazon.com and other booksellers. You can also register for our online courses or study groups of the same names. Learn more at FamilyLocket.com/services. To share your progress and ask questions, join our private Facebook group by sending us your book receipt or joining our courses to get updates in your email inbox each Monday, subscribe to our newsletter at FamilyLocket.com/newsletter. Please subscribe, rate and review our podcast. We read each review and are so thankful for them. We hope you’ll start now to Research Like a Pro.