Have you used Google’s AI tool, Gemini?
I’m still learning how to use Gemini and where it might fit into my genealogy research. Recently, I decided to test it with a Civil War question involving one of my ancestors, Noah Briles, who served in Company I of the 1st Iowa Cavalry.
I started with a broad prompt:
“Research the 1st Iowa Cavalry during the Civil War.”
I then narrowed the request:
“Trace Company I of the 1st Iowa Cavalry.”
Those searches helped me learn more about the regiment’s engagements and movements during the war. I also learned that many of the soldiers in Company I came from Wapello and Keokuk counties in Iowa—useful geographic context for understanding the men who served alongside Noah Briles.
Then Gemini offered to help look for an ancestor’s muster records. So I tried another prompt:
“Research Noah Briles in Company I of the 1st Iowa Cavalry.”
That was when the search became especially interesting.
Curious about the source behind some of the information in Gemini’s response, I followed a link to Wikisource. The link took me to a digitized volume of the United States Statutes at Large for the second session of the 65th Congress in 1918.
There, I found the following entry:
The name of Sarah J. Briles, former widow of Noah W. Briles, late of Company I, First Regiment Iowa Volunteer Cavalry,and pay her a pension at the rate of $25 per month.
That was a source I had not previously encountered.
The wording identifies Sarah J. Briles as the former widow of Noah W. Briles and directly connects him with Company I, First Regiment Iowa Volunteer Cavalry. It also documents congressional action involving her pension at a rate of $25 per month.
Not all of the information Gemini provided about Noah Briles was accompanied by sources. However, much of it aligns with information I already have in my public Ancestry tree, in previous blog posts, and in entries for Noah Briles on other genealogy websites. That agreement does not make the AI response a source, but it does help identify clues that can be checked against actual records.
And that may be the most valuable part of this experiment.
A simple series of prompts led me from a broad question about a Civil War regiment, to a specific company, to an individual ancestor—and finally to a federal source I had not previously used.
Even better, the discovery suggests a new research possibility. If a private pension act for Sarah J. Briles appears in the United States Statutes at Large, could similar entries exist for other ancestors or their widows?
That is a question worth exploring.
For me, this Gemini experiment was a reminder that AI does not have to provide the final answer to be useful in genealogy. Sometimes its greatest value is helping us discover a source we did not know existed.
And in this case, one simple search connected me with a new source—a source that may have great potential for researching other ancestors.