When starting the Accreditation process through the International Commission for the Accreditation of Professional Genealogists (ICAPGen), the first step is to decide on a regional focus and then find four connecting generations to research for the Four-Generation Project. You’ll need to analyze your pedigree to find suitable candidates. Artificial Intelligence can be a valuable tool in this process, helping to overcome common challenges and providing new perspectives on your family tree data.
Challenges with Pedigree Analysis that AI Can Help Overcome
A challenge in choosing a Four-Generation Project is to locate four generations that meet the specific parameters for accreditation, saving hours of manual searching. If you haven’t settled on a testing region, AI can provide several possibilities that you can evaluate. You will want to consider the research already conducted, available records, and your interest in the region.
Although I have ancestors throughout the US South, when I accredited, I chose the Gulf South Region – now split into Southeast and South Central. I had already completed the bulk of my Royston research and felt confident in producing the needed evidence. But could I have chosen another region? If I want to pursue a second accreditation, what would be the possibilities? I decided to try finding possible four-generation accreditation projects using AI.
Working with AI Tools:
There are several ways to approach pedigree analysis with AI, each offering unique advantages for genealogical research. The key is understanding how to effectively use these tools to enhance our research process. I used two models, ChatGPT and Claude, to undertake this task.
Beginning with ChatGPT, I asked if it could analyze a family tree, and it replied that it could use a GEDCOM file, CSV, or a text document. I then asked how large a GEDCOM it could handle, and it replied up to 100MB. If larger, the GEDCOM could be split into smaller parts.
ChatGPT has custom GPTs that are free for all users. They allow anyone with a Pro account to upload documents and set custom parameters for a chat. You’ll find the GPTs on the left of ChatGPT under “Explore GPTs.” You can search by any topic of your choice. By searching for “family tree,” I discovered the FamilyTree Expert GPT, where you can upload a GEDCOM and ask the GPT to analyze it.
1. Creating a GEDCOM File:
Preparing your data for AI analysis requires careful consideration of what information to include. When exporting your GEDCOM file, it’s important to be selective about the number of generations you include, as too much data can overwhelm the analysis. At the same time, too little might miss important patterns. Select your starting person thoughtfully, considering their position in your family tree and the generations you want to analyze. Include basic facts that will be most relevant to your research goals, but be mindful not to include unnecessary details that might confuse the analysis. Consider what additional notes or details might provide context to help AI better understand your research needs.
Because my paternal line has various locations and is colonial US, I created a GEDCOM in my FamilyTree Maker software with my father as the starting point and going back ten generations. I exported only the basic birth, marriage, and death facts to simplify the AI’s analysis. The four-generation project requires research on four individuals who have lived in the chosen region for at least one life event, so this should provide enough information. For more information on the specifics of choosing a four-generation project, be sure to reference the Guide to Applying for an Accredited Genealogist Credential.
2. Using AI for Analysis:
When working with AI tools for pedigree analysis, clarity and specificity are essential. Provide clear parameters about what you’re looking for, whether it’s specific time periods, locations, or types of records. When searching for candidates for a four-generation project, be explicit about geographical regions of interest. Request analysis of life events in particular locations to ensure your candidate meets accreditation requirements. The AI can help identify patterns in migrations and family movements that might make certain lines more or less suitable for your project.
I asked AI to find four individuals with life events from the Upper South region of Kentucky, North Carolina, Tennessee, Virginia, and West Virginia. It provided four linked generations that lived in North Carolina or Virginia. However, this would be a challenging project given the date range of 1708-1860. ICAPGen has defined the starting individual as one who needs to be born 80 years before the current year, so an easier project would be between 1850 and 1940.
3. Refining Your Search:
You’ll need to refine your search parameters carefully to get the most useful results from AI analysis. Be specific about the time periods you’re interested in, as this helps the AI focus on relevant generations. Include location requirements that align with your accreditation region. You could even ask the AI to analyze gaps or inconsistencies in your data, as these might indicate areas needing additional research before proceeding with your project.
I wanted to use another AI model, Claude, for this experiment using a different format for my pedigree. Within my genealogy software program, I created an Ahentafel chart of those same ten generations, beginning with my dad, that I uploaded to Claude for analysis. I used the same prompt about finding four individuals from the Upper South region.
I decided this time to ask for individuals in the South Central US region, which includes Texas and Oklahoma. I have plenty of ancestors to choose from in that region and was curious to see what Claude would find. Claude struggled with this task and needed several prompts to refine the results. Errors included skipping a generation and providing two individuals in the same generation. It acknowledged that it wasn’t accurate but noted that it had the correct locations.
After a few prompts refining the results, Claude finally identified an accurate four-generation linkage.
Once the generations were correctly identified, I prompted Claude to create a CSV file that I could upload to a spreadsheet and save. If I wanted to accredit in the South Central US region, I would now have the specific lines outlined and be well on my way for my four-generation project.
Best Practices:
When using AI for pedigree analysis, it’s crucial to maintain high research standards. Always verify any suggestions or insights provided by AI against your existing records and documentation. I was surprised to see how AI became confused when looking at the generations. The responses sound so confident! Before making your final selection for your four-generation project, consider multiple candidates suggested by the AI analysis, evaluating each against the project requirements. Although AI found two possible paths for a four-generation project, the Texas-Oklahoma one would be much more doable than the Virginia-North Carolina project.
AI can be a powerful assistant in analyzing your pedigree for a four-generation project, helping you identify promising candidates and potential research challenges. However, it’s important to remember that AI is a tool to enhance your research process, not replace traditional genealogical methods. By combining AI analysis with solid research practices, you can more efficiently identify and evaluate potential candidates for your four-generation project.
Best of luck in all your genealogical endeavors!
2 Comments
Leave your reply.