TECHNOLOGY TUESDAY ~ Using AI to transcribe old probate records

Image created by ChatGPT

We all are aware that AI is and has been being used in genealogy for some time now. 

Whether you’ve embraced it or avoided it, is a personal decision.

I like to explore new technologies as they are available and have been tinkering with AI for about a year now. 

What I’ve been doing is nothing compared to other genealogists who have jumped in whole hog with both feet. A good example of that would be our friend and blogging guru, Randy Seaver of Geneamusings.  If you want to see what he has been experimenting with I encourage you to visit his blog. He has an entire section/tab devoted to this subject.

For my part I have used it to transcribe some obituaries and probate records and to create a few images. I used both ChatGPT and CoPilot for those tasks. With excellent success.

This week I used ChatGPT to transcribe a handwritten probate from 1745, for my paternal 7^th great grandfather, Richard Everett. Please click here to see a post about this ancestor.

Here is what the probate record looks like.

(Please click on any image to enlarge it)

First 2 pages of probate record for Richard Everett - 1745

What I found was that I’m getting better at using this tool.  Better at asking the questions in a way that gets me the results I am looking for.

I began by uploading the first double page of this probate and asking “please transcribe this document exactly.”

ChatGPT asked a question about layout, I answered and got my transcription. 

However, as I read through it (which is critical so that we can catch errors or hallucinations).

General definition of hallucinations when referring to AI  

“AI hallucinations are incorrect or misleading results that AI models generate. These errors can be caused by a variety of factors, including insufficient training data, incorrect assumptions made by the model, or biases in the data used to train the model.”

There was a phrase that AI used over and over in the transcription, that did not appear in the original document.  I then asked this – 

“can you please re read this document and remove the words "sheeps for life" as that phrase does not appear in the document. Try to read the document again and do a better job of transcription”

Much to my surprise (because I haven’t done this process a lot) Chat did what I asked and the next version was nearly perfect.

I was then asked if I wanted a full cleaned transcript of all pages as a continuous document, including OCR that is searchable.

I answered that I would love that.

Here is a portion of the text from the image above, created by ChatGPT.  I found a couple of errors and will make the corrections as necessary.  But, what a time saver this is for us.

That was done and then I was asked – “Would you like me to create a genealogical abstract and analysis next — summarizing heirs, relationships, property, and inferred family structure (standard NEHGS-style abstract)? That would make it ready for citation or inclusion in a compiled genealogy.”

Again I said yes.  And I was truly amazed at the summary Chat created for me. 

This entire analysis is 5 pages. But, what a way to summarize what was in this will, including relationships.  

I'm so impressed!

🕮 Genealogical Abstract & Analysis

CONCLUSION – I am convinced that I will continue to use AI as my virtual assistant in my research. I hope you will give it a try.  And, please do let me know if you do.

Happy hunting,

Michigan Girl
Copyright ©  2010-2025   Diane Gould Hall
ALL RIGHTS RESERVED.  NO USE WITHOUT PERMISSION

 

FRIDAY FINDS ~ Paternal 7th great grandparents - the Everett family of Massachusetts

 

Doing a bit of follow up to some research I did while I was at Rootstech and the Family History Library, this past March.

I don’t know about you, but it generally takes me quite a while to get back to some of this research I do while I was at the FHL.

One family I found information on while I as there, was the Richard Everett family of Dedham, Massachusetts.  This family is linked to my Mayflower line, and may lead me to another Mayflower line, yet to be determined.

In this case I am trying to prove the marriage & children of Richard Everett and his wife, Mary Fuller.  They would be my paternal 7^th great grandparents. And, I’d like to have more proof of their relationship and the children they had together.

NOTE: When we are researching our ancestors back in the 1600’s we must do a bit of digging.  Because we all know just because someone else’s tree says it’s so, doesn’t mean we believe it.

What have I located so far, in the way of proof?

Let’s begin with the birth record for Richard Everet (also spelled Everett) in 1683 in Dedham, Colonial Massachusetts.

(Please click on any image to enlarge it) 

Birth record - Richard Everit, son of John and Elizabeth born 24 Aug 1683

Now, how about a marriage record for Richard & his wife, Mary Fuller.

Marriage record - Richard Everet to Mary Fuller - 3 Mar 1708 in Massachusetts

NOTE: These are images of the actual records and not an index that has been interpreted by a transcriber. It’s important to find the originals whenever we can.

Here’s another source with names, dates and places.  Remember these kinds of documents are submitted with proof, however, the proof is not provided with these images. It is a document/application submitted by someone.  However, what it does give us is leads, that combined with other evidence allows us to draw a conclusion.

Application for Sons of the American Revolution provided by a descendant of Richard Everett and Mary Fuller

Here’s another reference to the marriage of Richard & Mary.

U.S. and International Marriage Records, 1560-1900

Name	Mary Fuller
Gender	Female
Spouse Name	Richard Everett
Spouse Birth Place	MA
Spouse Birth Year	1683
Marriage Year	1708
Number Pages	1

One of my favorite sources are wills/probate records.  When our ancestors were dying or suspected they were nearing that age, they often made wills.  It was important to them that their spouses, children and any other heirs were taken care of.  These can be gold mines of information, including married names for daughters, locations of living descendants, names of other family members etc.

Here then, is the will of Richard Everett dated 1745/6 in Dedham, Suffolk County, Massachusetts. 

In it he names his “dearly beloved wife” Mary and his sons, Joshua, Jeremiah, Jonathan, Israel and Timothy. He later mentions his “beloved” daughter Mary Maxey, “beloved” daughter Bethiah Day, “beloved’ daughters Hannah & Abigail Everett.

What I learned from this document, was that there were 2 sons and 3 daughters I had not yet discovered for this couple. More research into these additional children will be done.

And, I went back to the Massachusetts Town & Vital Records and located the births of Timothy, Jonathan, Isarael, Bethiah born 18 Nov 1721 and another Bethiah born 9 October 1722 and Hannah.  Because there are two daughters named Bethiah, born within a year of one another, I will look for a death of the first one.

In conclusion.  

• Have I found enough evidence to prove that Richard Everett married Mary Fuller? 
• And that they had 5 sons and 3 or 4 daughters? Yes, I believe I have.

Do you have any input? Would you have come to a different conclusion?

Sources:

¹Massachusetts, U.S. Town & Vital Records 1620-1988

² U.S. and International Marriage Records, 1560-1900

³ U.S., Sons of the American Revolution Membership Applications, 1889-1970

⁴ Massachusetts, U.S., Wills and Probate Records, 1635-1991

⁵ FindAGrave memorial for Mary Fuller Everett #64356211

⁶ Massachusetts, U.S., Compiled Birth, Marriage, and Death Records, 1700-1850

If you are related to or connected to anyone in this blog post, please get in touch.  Let’s exchange information.

Happy hunting,

Michigan Girl

Copyright ©  2010-2025   Diane Gould Hall

ALL RIGHTS RESERVED.  NO USE WITHOUT PERMISSION