In this episode of the Research Like a Pro Genealogy podcast, Diana and Nicole discuss using AI in timeline creation and source analysis. Diana shares her experience using AI tools to transcribe, abstract, summarize, and analyze documents for her Isabella Weatherford project. She explains how AI can be a powerful ally in genealogical research, but emphasizes the importance of verifying AI-generated information and using your own genealogical expertise to interpret the results.
Diana used AI to transcribe Isabella’s 57-page pension application, extract key events and dates, and create a structured timeline. She shares the challenges she faced and the different AI models she tried, ultimately finding a combination of Transkribus, ChatGPT, and Claude to be the most effective. Diana also demonstrates how she used AI to analyze the sources and information within them, teaching the AI model to differentiate between original, derivative, and authored sources, as well as primary, secondary, and undetermined information.
Listeners will learn how to use AI tools to streamline their research process and gain valuable insights from their ancestors’ documents. They will also learn the importance of maintaining a critical eye and not relying solely on AI-generated results.
This summary was generated by Google Gemini.
Transcript
Nicole (1s):
This is Research Like a Pro episode 324, Using AI for Timelines and Analysis. 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 (42s):
Today’s episode is sponsored by Newspapers.com. Hi everybody. Welcome to Research Like a Pro today. And hi Mom.
Diana (50s):
Hi Nicole. how are you doing?
Nicole (52s):
I’m great. I’ve been up really early this morning working on finishing up a Dyer research report. I wanna finish that one up before I get started on the next one for the study group, so I’m a little behind.
Diana (1m 7s):
That’s a good motivation to get that one finished.
Nicole (1m 11s):
Yeah,
Diana (1m 11s):
Well what have you been working on?
Nicole (1m 14s):
Well, I am trying to finish that Dyer report and then also since we had our first lesson in the Research Like a Pro study group yesterday and I was teaching about objectives and we are incorporating AI tools into these lessons this time I was thinking about making a custom GPT that helps evaluate your research objective to see if it’s too broad or too narrow, if you’ve included too much information, or not enough unique identifying information. And also to help you break your research into phases if you have a complex research question. So I published that and put it in the GPT store and hopefully that will be a good way to put in your research objective and get some feedback.
Diana (1m 54s):
Oh, what a great idea. What did you name it so we can find it?
Nicole (1m 58s):
Genealogy Research Objectives.
Diana (1m 60s):
Okay, well that is pretty clear.
Nicole (2m 3s):
It’s very generic sounding, not very creative, but hey.
Diana (2m 6s):
Nope. Hey, that’s, that’s good. Well I am excited to try that out because I am really trying to figure out what objective I want to work on for the research project. I have so many different things I could do going forward from my Cline and Weatherford project from last time and, I’m just not sure if I wanna try to trace John Cline or find candidates for him in 1830. You know, there’s so many John Clines, or do I want to maybe trace Henderson Weatherford’s parents and try to confirm his parents and his family? Hmm. There are so many WeatherFords as well, so I am just conflicted a bit. I’m gonna have to do some real thinking about what I wanna do for that objective.
Nicole (2m 50s):
Yeah, that’s tricky because you wanna pick the one that leads to the answer, but you can’t predict which one will be right, the right path. You’ll probably have to do both.
Diana (3m 0s):
I will definitely have to do both. And I’m kind of leaning toward Henderson Weatherford because everything pointing to his parents that everybody has accepted on this line forever and ever is indirect evidence. And I want to shore that up in my own head. I wanna make sure that I feel like that is absolutely correct. So I, I probably will do that and I’m thinking there might be a Weatherford-Cline connection back in Illinois and Kentucky and so I’m thinking, well if I can give the Weatherford line straight, maybe I’ll happen into the Clines, you know, so maybe that’s what I’m going to do
Nicole (3m 38s):
For sure. Because I mean sometimes the people who met and got married lived in the same town and that could lead you to the right Cline family.
Diana (3m 47s):
I know. Well it’s always good to talk these things out because then you get it clear in your head.
Nicole (3m 53s):
Yeah. Well let’s do some announcements. The Research Like a Pro Webinar Series for October is going to be with Emma Lowe, one of our Family Locket Genealogist researchers. She’ll be discussing her research on Crossing the Pond: Tracing Dorothea’s Roots in 19th Century Pomerania. And this will be about a passenger list recorded by a particularly conscientious clerk and a new index entry for a German civil registration record provided clues that opened the doors to tracing Henry and Dorothy Trotz’s origins in Greifswald, Pomerania, Prussia, Germany.
Nicole (4m 40s):
This case study follows the Research Like a Pro process to systematically identify and search relevant record collections, following the clues that enabled the discovery of new generations in the homeland of Germany. The topics shall cover include Germany, immigration, Evangelical-Lutheran church records, German civil registration, locality survey, resolving conflicting evidence, German research resources. Emma is a professional genealogist and she’s a great member of our team. She is pursuing a Master’s degree at the University of Strathclyde in Genealogical, Paleographic, and Heraldic Studies. She has a Bachelor’s degree in Family History with an emphasis in German-American immigrant research from Brigham Young University and she also enjoys British, African-American, and DNA research.
Nicole (5m 28s):
So we are excited to hear from Emma on October 15th and if you haven’t registered yet for our 2024 webinar series, you can do so on our website and the cost is $60 and you get access to the 12 webinars for the year. Alright, well we just started our Research Like a Pro Study Group for this fall and so our next study group will begin in February of 2025, Research Like a Pro with DNA and the registration for that will be begin in December. To be a peer group leader for that just apply on our website and send us in one of your DNA research reports. Join our newsletter to get all these news and updates and notifications of any sales and coupons we’re running. And upcoming conferences include East Coast Genetic Genealogy Conference. We’ll be speaking virtually at that in a prerecorded lectures.
Nicole (6m 9s):
And then Diana will be at the Texas State Genealogical Society Family History Conference, which is virtual November 1st and 2nd giving several lectures.
Diana (6m 18s):
All right, well exciting things coming up in the fall. We have a listener spotlight today and our listener is Mark Thompson, our friend. He’s also the co-host of the Family History AI Show, and Mark says, “Diana and Nicole, I wanted to share how much I enjoy today’s RLP podcast, episode 311. I found it to be both an in-depth and accessible discussion of AI disclosure, a topic that has many in the family history community stuck wondering what to do or where to turn for advice. Thanks for everything you two do for the community.” So thanks for that Mark. We appreciate that and we’re excited that you are doing a podcast as well on AI.
Diana (6m 60s):
Well, let’s get started on our topic for the day, which is a return to using the Research Like a Pro process with AI, and I’m sharing how I did this as I was working through a project on Isabella Weatherford and today we’re going to talk all about the timeline creation and the source analysis part of the process. So in the first episode about this, I talked all about using AI to explore research questions and writing the research objective. So let me remind all of you listeners what the research objective was that I finally settled on after I used some AI to help me.
Diana (7m 40s):
The objective of this research phase is to examine the economic and social conditions in Dallas County, Texas in the early 1870s and their influence on Isabella D Weatherford’s life and marriage prospects. Isabella was born on 4 March, 1858 in Missouri. She first married John H. Carpenter on 16 January, 1874 in Dallas County, Texas, then later married Robert Sidney Royston on 16 January, 1877 in Van Zant County, Texas and she died on 9 May, 1942 and to Tucumcari County New Mexico. So in this episode we’re going to talk about the AI tools and how I use those to create the timeline and analyze the sources.
Diana (8m 28s):
So the different models that I used for this project were Transkribus’ Text Titan 1, which is the supermodel, ChatGPT-4o which is Omni, and Claude 3.5 sonnet.
Nicole (8m 42s):
Great. You’re using all these great AI models and tools to help you with this. Well, artificial intelligence can be a good ally in our genealogical research and sometimes we face a lot of challenges when working with documents about our ancestors. When we have to transcribe, abstract, summarize, and analyze, these are areas where AI really shines. It takes an existing document and the information it holds and then transforms it for us to use as we assemble the timeline. So let’s talk about how you used AI to build Isabella’s timeline. So you started by gathering the data, collecting all the documents and information you had about Isabella, including census records, marriage certificates, her widow’s pension file, and then after that you realized you needed to transcribe the 57 page pension file, which I know you didn’t do because of how much work that would be.
Diana (9m 40s):
That’s right. I had read through it and pulled out some key pieces of information like what Isabella reported for her marriage date and civil war information about her husband. But man, that 57 pension page pension file was just daunting, so much in it.
Nicole (9m 58s):
Right? And that’s one of the great uses of Transkribus is that you can, you can use it to help you find things within a big file like that. Even if you don’t care to have it perfectly transcribed, you can still get a a searchable document, which is kind of what FamilySearch’s AI tool does. It makes all of these previously unindexed deeds and wills and notarial records searchable. Even though the transcription isn’t perfect, you can still find what you need in it a lot of the time. Well you wanted to make your 57 page pension file not only searchable but you actually wanted to have a good transcription of it. So you were going above and beyond just getting it searchable.
Nicole (10m 40s):
So you uploaded all those images to Transkribus and to ChatGPT and to Claude and you asked the large language models to transcribe it saying, “you are an expert genealogist, please transcribe this document accurately, preserving all names, dates and places exactly as written.” I love that prompt because it’s so hopeful that you, a large language model can do it. And we know that this is kind of a challenge for the AI tools and they use their computer vision to read the handwriting, but they’re getting better and they work especially well on consistent handwriting that you know, doesn’t have different changes in forms and things.
Nicole (11m 29s):
So what’s great is that you were able to get some results using these tools combined. So I’m excited to hear you tell us about that. Just a little more background about that is that it, it was a challenge because it had a variety of documents and it not only spanned 58 years, but it included all different types of documents, handwritten letters, forms where some things are printed, some are typed, some are handwritten, all kinds of different pieces of paper and different people writing on it with handwriting. So the file included all kinds of names, dates and places and different details about her life.
Nicole (12m 12s):
And so transcribing it fully really was a good idea to extract everything you could out of it.
Diana (12m 19s):
Yeah, it absolutely was. And it was so interesting to do the comparison. So I’ll just make a note that in Transkribus, you don’t give it a prompt it you just say, you know, do the transcription. And so I used that prompt with ChatGPT and Claude, but not with Transkribus, it just did its thing once I got it started. So first I decided I would try Transkribus because I could put everything in all at once. I had previously broken up this huge PDF into three parts because I was uploading it to FamilySearch and it was too big to put it all in at once. So I put my three parts in and then you know, you wait a while for Transkribus to do its thing.
Diana (13m 0s):
But when I looked at it I was pretty disappointed because it identified the lines correctly, you know, it always goes through and organizes it by line, but man, some of the things it just could, I mean it really didn’t do very well. It really struggled with most of them. So I was a little disappointed that it just didn’t come out perfect. You know, we’re always hopeful. So next I tried ChatGPT and Claude, but I didn’t upload the entire PDF because that would’ve been way overwhelming for the model to do. So when I went to ChatGPT or Claude, I kind of went back and forth testing them, I would just put one document, you know, I just would put a, a screenshot of a or an image, whatever I was using, put that right into it.
Diana (13m 44s):
And I found that they were actually pretty accurate for a lot of the things, but one of the ones that was really a problem was the actual pension application because it had was a typed form. I mean it was a pre-printed form with lines for the answers and then someone had gone in and typed in the answers over the lines and then there were a few things handwritten. So that one just came out to be a mess in both ChatGPT and Claude. But then I decided to see if I could take what Transkribus did because it had organized it better, And I put that in ’cause it had the correct line breaks.
Diana (14m 25s):
And then I asked ChatGPT or Claude to use the original as well as Transkribus’ version and then to fix it up. And that was amazing. That one worked so well. But the funny thing was that both of those models, ChatGPT and Claude, hallucinated and they just completely made up a few things. And so I had to really check it carefully, correct it until I got the correct version. But it’s so interesting how even though it’s clearly something, you know, a date and it was typed written so it was easy to read, still just hallucinated. So you know, it’s such a good learning opportunity to do your double checking and realize that you’ve got to be the human in the loop as we always say, that checks everything.
Diana (15m 10s):
So since I was transcribing one image at a time, and I sometimes used Transkribus to help ChatGPT or Claude to get the form right, but sometimes I just used one of the models. And I decided I needed to be organized, so I created a Google Doc, and I pasted a copy of the image, you know, I snipped the image, and I gave it a title. You know, I had so many things like letters and envelopes and forms. So I put what part it was, ’cause I had this three part PDF, I put the page number, the description, you know, kind of gave my identifiers for it. So I had the identifiers in bold, and I made them a heading, so they would show up with my headings in the outline on Google Doc, and then I pasted in the image and then I added the correct transcription, once I finally had it all correct, then I put that right below.
Diana (16m 4s):
So then I could see the original, And I could compare it. But I also had the transcription so I could use tools like control find to quickly find names, dates, places, and to use it really easily. So that’s something that I’ve developed over time that I did the Google Doc, I did it first with a land grant record from Texas, you know, doing the same thing with images and transcriptions. So it worked so well for this pension.
Nicole (16m 33s):
That’s great. I know a lot of people are wondering if you had done this on your own, would it have been faster? Do you think that this saved time for you? What are your thoughts?
Diana (16m 44s):
Absolutely. It saves so much time because there was a lot of text. I would’ve just been typing and typing. And. I feel like my typing isn’t as accurate as I would like. I can type fast, but I make a lot of mistakes. Sometimes I’ve gotten lazy because the AI that works with my typing, you know how it will just autocorrect and sometimes it’s inaccurate. So I would’ve had to check my own typing anyway against the record. So I felt like it did save me time and you know, it was so much more fun than actually doing it myself. I just thought it was fun to see how AI would do and it was way more fun. I think it only took me three to four hours, which still sounds like a lot, but for 57 images I feel like that saved me some time
Nicole (17m 33s):
For sure. Yeah, sometimes when I’m transcribing long documents like that, I get kind of sleepy just sitting there typing And, I can like nod off in the middle of it, but it can be helpful to have something that you’re editing. It’s a little more active than just like trudging through. So I like the idea that you can get something to work from that you pair and then you can notice differences and not have to generate the whole transcription yourself.
Diana (18m 1s):
Right. For me, I felt like it was a great process and I’ve done so much transcription in the past, I didn’t feel like I needed to practice those skills here. You know, I know how to do it. I just wanted to have a coworker work with me on this.
Nicole (18m 16s):
Yeah, that’s great. Well I just, I look forward to these tools improving with their computer vision. And I really think that Transkribus is a great place to work with because they can incorporate a language model and so I think the supermodel that you used was a good one. And I think that will get stronger and better. And I also think that both of the tools that you mentioned, Transkribus, which is focused on handwritten text recognition, as well as ChatGPT and Claude, which already have the language model, but they have the computer vision too. I think that transcript and the language models are going to get both better and kind of come to like a similar way of doing it.
Nicole (19m 2s):
That will work really well and if we kind of learn the best prompts and the best way to do it, then I think we can actually make this transcription of handwritten text really useful for us as Genealogists, even though it’s still kind of in the early days for it.
Diana (19m 20s):
I agree, I think it’s good to jump in and try it so we can see the developments, we can see how it improves and we’ll down the road we’ll say, remember when we put something in there and it hallucinated and we had to really check it, you know, hopefully we’re going to remember that and say now it’s so much better. So it’ll be fun to see how things evolve.
Nicole (19m 43s):
Right? 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. They’re 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.
Nicole (20m 29s):
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 (20m 43s):
Alright, so let’s talk about extracting information. So another reason I wanted to do this full transcription was to be able to extract all the key events and dates from the entire text. And I wanted to be able to ask the AI to do this for me. So I knew that with my Google doc having the images and the text, it would probably be too much information for the large language model to handle. It might confuse it. So I did a duplicate of the Google Doc And I deleted all the images. So it was just the text, just the complete transcription. And I felt like that would make it easier for the a AI model to extract information.
Diana (21m 26s):
I saved that Google Doc as a PDF file. And I uploaded it to chat GPT with this prompt, I said, “please create a chronological list of events mentioned in this transcription, including dates and places where available.” Then I used AI to convert the extracted information into a structured format. So not only did I want this list, I wanted to be able to put it into a spreadsheet and so I gave the next prompt create a CSV file with columns for date, place, person, event or activity using the information from the previous list. So this was a multi-step process here.
Diana (22m 8s):
First I asked it to extract all the information and make it chronological, and then I asked it to create the CSV file, which is comma separated file.
Nicole (22m 18s):
Yeah, that’s good. Let’s see. What is a sample of what this AI generated timeline looks like in the CSV format? Well, there’s four columns. So date, event, place and source. And then looking at it with the comma separation, you would first see a date like “1858, birth, Springfield Missouri, Pension application”. So that’s what it would look like in the CSV format,
Diana (22m 50s):
Right? And the thing that I appreciated was I didn’t really know how to get that CSV format into my spreadsheet. I wasn’t sure exactly what to do, but it gave me instructions. It said, put this into, copy this into Notepad, and then from Notepad, save it as Excel. It’s like, oh, Thank you for helping me out with that. So it was really an interesting exercise there. So I noticed when it gave me the information that it, I had forgotten to ask it, the specific data point where it would get, say the date of the marriage or the death. So then I said, I went back, And I asked it to give a column for the reference and I also asked the AI to change the date format to be consistent with day, month, year, like we’d like to do in our genealogy research.
Diana (23m 41s):
Because first of all, it was giving me the year and then the month and then the date probably because it was maybe going chronological. I don’t know, it was just giving it to me different. So I had to keep asking it to get it exactly how I wanted it. But I really liked it when I got the final Excel spreadsheet because I had all the dates, places, persons event, and, I had the reference. So I had the reference that I had put into my Google doc, like this is part one, this is page three, this is a note, whatever it was. And so I could see, oh, that’s where the marriage was and that’s exactly where it got that information. So with everything really clearly laid out in the spreadsheet, it made it so I could add all of the pertinent events to my Airtable research log.
Diana (24m 30s):
And I was using the 2024 template. So first of all, I created the entry for the overall pension file in the research log table of the Airtable template. And then as I was creating my timeline, and I select, I would select that pension as the source for each event. And then I copied and pasted the transcription from my Google doc straight into the event details field. And we can expand that, you know, to put as much information as we want. So I could put a full transcription for each event, which was great. And then I was able to add a note to the specific detail, like the part page and the information.
Diana (25m 14s):
I just added that in my notes field because my source citation is for the whole pension. You know, I don’t, I’m not going to go do that right now to add that specific piece to my citation. I will do that in a research report, but then I have that little detail to add to my citation already for when I want to write my report or if I want to add, you know, that specific little piece, upload that to FamilySearch or Ancestry. So it was kind of, I had to really think through what I wanted to use this for and how I wanted to do it. So just some tips that I discovered for transcribing documents for your timeline that as you saw that I tried multiple models and you need a good prompt.
Diana (26m 4s):
And then Steve Little taught us in our, in our Institute class, and he teaches in his webinars and lessons that we should give the model this kind of a prompt. We give it the role. So what is it? You’re an expert genealogist. We give it the goal, you’re supposed to transcribe this record and then we provide the text for it. And then we tell it, tell what the task is. We want you to do this, what the actual names and dates and places. And then we can describe a flask like I want the some bullet points or I want it in the CSV file, however you want that. After you do your prompt, then you check your names, dates, and places very, very carefully and you just keep refining until the result is accurate.
Diana (26m 51s):
You know, you just keep saying, no, that’s wrong. Here’s the correct one. Get it. Get it how you want it. And I think it’s so important that we just use our imagination to think of new and better prompts. And we’re really not limited in what we ask it to do. We just need to think through what we might want it to do.
Nicole (27m 8s):
Great tips. Well, an important part of the research process is to analyze each source and piece of information that’s in the timeline, especially as you’re getting started to see how reliable your information is. So you were curious to see if AI could help with it. So you asked it to describe original derivative and authored sources according to genealogy experts. And its answer was pretty good, but a few things weren’t right and so you could correct it. And that’s the beauty of having a chat with the large language model is that it learns from within the conversation and remembers what you tell it. So for example, as Genealogists, we recognize that an image copy can still be classified as an original record, not as a derivative.
Nicole (27m 56s):
So when we look at images online, that’s still an original image. But ChatGPT didn’t know that,
Diana (28m 4s):
Right? It thought it was a derivative. So I had to correct that.
Nicole (28m 8s):
And then as you were talking more about the source analysis with the chat bot, it corrected itself, so it learned. And when you were happy with the analysis, you could copy and paste that directly into your timeline into the column about source analysis. And so this kinda saved you some time writing out the source analysis yourself. And that’s just one of the really great things about language models is their ability to write coherent sentences. And if we like the way it sounds and it’s the way we would’ve written it, we can use it and save us some time and fill out our timeline
Diana (28m 47s):
Right. It was really fun to do it this way. I’ve done the source analysis for years and you know, usually I’m pretty brief in my reasoning. I’ll just say, yeah, this is an original or this is a photocopy of a of original marriage record or, or whatever. But the large language model went into quite a bit more detail and it was kind of fun to see the different things that it came up with that I agreed with. I just would not have written out that much. So it was nice to have something really complete to put into my timeline. So another thing I wanted to see if it could analyze information within a source. So I then asked the model ChatGPT to describe information based on primary, secondary, and undetermined information.
Diana (29m 34s):
And it got confused on this. So I had to explain that the source is the container for the information and there could be various pieces of information in one source. So I tested it by uploading a death certificate, which we know as Genealogists will have all sorts of pieces of information and all sorts of different informants and ChatGPT recognized there was both primary and secondary information and identified the informants, gave me details and its reasoning. And again, some of the names and dates weren’t correct. It couldn’t quite read the doctor’s handwriting, so I had to give it feedback and it just misinterpreted a couple things but I was very happy with the finished product.
Diana (30m 19s):
And I added the information analysis to my timeline as well. So just like you created a custom GPT for objectives, I decided to create a custom GPT on analyzing sources. And I titled it Diana’s Genealogy Source Experts. So all of our listeners can go to ChatGPT and try that out. And all you do is copy and paste your document into the chat and then say, analyze this document for the type of source and information and see what you think, you know, if it doesn’t quite do it right then, you know, teach it, tell it what it did wrong and see if you can get it to give you exactly what you want.
Diana (31m 0s):
So I know you tried it out Nicole, what did you think the the custom GPT?
Nicole (31m 4s):
Oh, I thought it was great. I think it’s a helpful way to get started with your analysis and get you thinking. And of course we need to recognize that it’s not always gonna be right, but it’s a good way to get started and we can compare it with what we know about source analysis and get some ideas.
Diana (31m 22s):
I think it was really helpful for me to get some new thoughts on it because sometimes I kind of zoom through this part of the process and just don’t maybe think through everything because I think I, I’ve done this a hundred times, but it gave me some food for thought and some of the things that it came up with and whether I agreed or not. So it’s like if you’re talking to a friend and saying, what do you think about this is this primary information? Who’s the informant? I think it really helps us to get our reasoning clear in our head and gives us, gives us something else to bounce our ideas off of.
Diana (32m 2s):
So you know, it’s kind of a fun thing to do. And I would definitely not use AI to replace our own learning about information. So if you’re just getting started learning how to do this, you’ll want to not believe everything it says. You’ll want to make sure you have a solid foundation and then you can use it to see if you can get some, get some additional ideas.
Nicole (32m 28s):
Well, let’s go over some tips for using AI with source analysis. The key is to have a good working understanding of how to analyze sources and information. And if you already understand the terms, then you can recognize any errors that the large language model makes. I noticed sometimes it would bring in the terms that historians use, like primary and secondary sources, but in genealogy we differentiate sources as original derivative and authored where the information we differentiate as primary and secondary. So it’s good to really understand those terms before you try to use the language model so that you can make sure you don’t get things mixed up.
Nicole (33m 11s):
And it’s good to use these models to check your assumptions. And then also we need to check the AI’s assumptions and maybe start with a document where you already understand the source and information analysis and, and so it’s not thinking for you, but it’s maybe just generating the writing for you that you can then put into your timeline. You know that you can have a paragraph about how this is an original record and it, it was an image of the original or whatever like that. Be sure to chat back and forth with the model to make the answers more reliable to give it feedback and corrections. And then just copy and paste the final answer into your timeline to speed up that process of analyzing all the sources and information that you start with in your project.
Nicole (34m 1s):
Well, it’s really cool how to see how AI tools helped you with streamlining the process of creating Isabella’s timeline and analyzing the sources. So good job with that. And so it’s good to remember that AI is a tool to assist our research but not replace our critical thinking. So we always need to verify AI generated information and then use your own genealogical expertise to interpret the results. And in our next episode, we are going to explore more about Dallas County in the 1860s and 1870s and the historical context and how life at that time might’ve influenced Isabella’s life choices and her first marriage and all of that.
Nicole (34m 44s):
So stay tuned as we continue to go through this case study about Isabella Weatherford and the AI tools that Diana used.
Diana (34m 54s):
All right, well thanks everyone for listening and we will talk to you next time. Bye-Bye
Nicole (35m 1s):
Bye. 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.
Links
Genealogy Research Objectives Custom GPT – https://chatgpt.com/g/g-X7g8E5KSc-genealogy-research-objectives
Using AI in Timeline Creation and Source Analysis: Isabella Weatherford Project Part 2 – https://familylocket.com/using-ai-in-timeline-creation-and-source-analysis-isabella-weatherford-project-part-2/
Diana’s Genealogy Source Expert Custom GPT – https://chatgpt.com/g/g-2N0HVSfLX-diana-s-genealogy-source-expert
Sponsor – Newspapers.com
For listeners of this podcast, Newspapers.com is offering new subscribers 20% off a Publisher Extra subscription so you can start exploring today. Just use the code “FamilyLocket” at checkout.
Research Like a Pro Resources
Airtable Universe – Nicole’s Airtable Templates – https://www.airtable.com/universe/creator/usrsBSDhwHyLNnP4O/nicole-dyer
Airtable Research Logs Quick Reference – by Nicole Dyer – https://familylocket.com/product-tag/airtable/
Research Like a Pro: A Genealogist’s Guide book by Diana Elder with Nicole Dyer on Amazon.com – https://amzn.to/2x0ku3d
14-Day Research Like a Pro Challenge Workbook – digital – https://familylocket.com/product/14-day-research-like-a-pro-challenge-workbook-digital-only/ and spiral bound – https://familylocket.com/product/14-day-research-like-a-pro-challenge-workbook-spiral-bound/
Research Like a Pro Webinar Series 2024 – monthly case study webinars including documentary evidence and many with DNA evidence – https://familylocket.com/product/research-like-a-pro-webinar-series-2024/
Research Like a Pro eCourse – independent study course – https://familylocket.com/product/research-like-a-pro-e-course/
RLP Study Group – upcoming group and email notification list – https://familylocket.com/services/research-like-a-pro-study-group/
Research Like a Pro with DNA Resources
Research Like a Pro with DNA: A Genealogist’s Guide to Finding and Confirming Ancestors with DNA Evidence book by Diana Elder, Nicole Dyer, and Robin Wirthlin – https://amzn.to/3gn0hKx
Research Like a Pro with DNA eCourse – independent study course – https://familylocket.com/product/research-like-a-pro-with-dna-ecourse/
RLP with DNA Study Group – upcoming group and email notification list – https://familylocket.com/services/research-like-a-pro-with-dna-study-group/
Thank you
Thanks for listening! We hope that you will share your thoughts about our podcast and help us out by doing the following:
Write a review on iTunes or Apple Podcasts. If you leave a review, we will read it on the podcast and answer any questions that you bring up in your review. Thank you!
Leave a comment in the comment or question in the comment section below.
Share the episode on Twitter, Facebook, or Pinterest.
Subscribe on iTunes, Stitcher, Google Podcasts, or your favorite podcast app.
Sign up for our newsletter to receive notifications of new episodes – https://familylocket.com/sign-up/
Check out this list of genealogy podcasts from Feedspot: Top 20 Genealogy Podcasts – https://blog.feedspot.com/genealogy_podcasts/
Leave a Reply
Thanks for the note!