09 September 2025

My ChatGPT Obituary Experiment

Disclosure: This post contains results from ChatGPT 5, The content is clearly identified and presented without edit.

I'm well aware AI is best used as a supplement to regular research. It can certainly do some nice things. I use it extensively for formatting, transcriptions, and even providing a list of the important genealogical facts from newspaper articles, wills, and obituaries (though I still double check it hasn't missed anything). I do not use AI for actual research. The vast majority of records are simply not available to AI. Many data sets require logins and quite often, subscriptions. But even if AI can't answer a question directly, it doesn't mean it can't be helpful in pointing me in a useful direction.

Just for fun, I recently decided to see if ChatGPT could help me locate an obituary. It had been some time since I requested this type of information and I was eager to see the improvements in results. FindAGrave contained the text of an obit for my granduncle, John Burton Leroy. As shown on FindAGrave:
Mr. Leroy was a son of John Leroy and his first wife, Cynthia Cunningham Leroy. John Burton Leroy was survived by his stepmother, Mrs. Addie Leroy, Dalton, Georgia; and his siblings, Ben Leroy, George Leroy, Bertha Henry and Eliza Fowler. 
I previously found and verified everything contained in this snippet, with the exception of one piece of data....his mother's maiden name being Cynthia Cunningham. I know her name was Cynthia from a census record, but I haven't found any other proof of her maiden name. If I could put eyes on the obit, and not just a transcription, I'd happily add this fact to my tree.

So I asked ChatGPT 5, "Can you find any public newspaper articles containing this exact wording 'John Burton Leroy died at his residence on Gadsden Route 4'" the phrase at the beginning of the transcribed obituary. I actually did this search two different times, a couple weeks apart just to see the differences.

For the first response, AI thought for 19 seconds, then replied:

Doris — I searched for that exact string and didn’t find any public newspaper page that matches it verbatim.

What I found

No hits for the quoted phrase “John Burton Leroy died at his residence on Gadsden Route 4” in open web results.

There are near-matches showing the style you’re expecting (Gadsden Times obits often read “Gadsden Route 4, died … at his residence”), e.g.:

“J.A. Pankey, 69, Gadsden Route 4, died … at his residence …” (Find A Grave transcription of a Gadsden Times obituary). (Find A Grave)

Similar phrasing appears in other Etowah County obit transcriptions noting “Gadsden Rt. 4 … died … at the residence.” (Find A Grave, Geni)

You likely have the right person/place: there’s a FamilySearch profile for John Burton LeRoy (1878–1949) tied to Gadsden/Etowah research, but it doesn’t quote an obituary. (FamilySearch)

Why you may not see an exact match

Obituary lines commonly read “at his residence, Gadsden Route 4” (comma, not “on”) or “Rt. 4 / “Rte. 4,” and the surname appears both as LeRoy and Leroy in indexes. Any of those small differences will prevent an exact-phrase hit.
 
Next steps I recommend

Try these exact-phrase variants (including punctuation/casing):

"John Burton LeRoy" "Gadsden Route 4"
"John Burton Leroy" "Gadsden Rt. 4"
"at his residence, Gadsden Route 4" Leroy
"LeRoy" "at his residence" "Gadsden"

Check the Gadsden Times around his death year (likely 1949). Most images are behind paywalls:

Newspapers.com (Gadsden Times) — strongest coverage.

Google News Archive sometimes has scattered issues.

Alabama Dept. of Archives & History and the Alabama Digital Archives indexes occasionally include Etowah items. Wikipedia

Local help: Etowah Historical Society or the Gadsden Public Library microfilm desk can do a lookup if you provide name + month/year of death.

If you want, tell me the exact death date you have for LeRoy and I’ll target the Gadsden Times issues for that week and craft precise queries (and, if you have access to Newspapers.com, I can give you copy-paste search strings to run there)
For my second attempt, AI thought a bit longer at 37 seconds, then replied:
Hi Doris — I searched for that **exact** wording and couldn’t find a publicly viewable newspaper page that contains the phrase “**John Burton Leroy died at his residence on Gadsden Route 4**”. Nothing matching that string (including the trailing double-apostrophes) turns up on the open web or in free Alabama newspaper repositories.

However, context strongly suggests the obituary you’re after would be in **The Gadsden Times** around **December 14–16, 1949**, which aligns with FamilySearch listing **John Burton LeRoy**’s death as **December 13, 1949** (Etowah Co./East Gadsden). ([FamilySearch][1])

### Where it’s most likely hiding (paywalled/onsite)

* **The Gadsden Times** archive (hosted on Newspapers.com / Gannett) for mid-December **1949**. The Times is the local daily that routinely printed “Route 4” addresses and “died at his residence…” language in that era. The paper’s history and coverage area are documented here. ([Wikipedia][2])

* **Northeast Alabama Genealogical Society (NEAGS)** at the Howard Gardner Nichols Memorial Library in Gadsden — they maintain obituary clippings and local indexes; staff/volunteers can usually pull a specific death notice by date. ([Wikipedia][3])

### How to zero in on the exact string

OCR in newspaper databases often renders **Route 4** as *Rt. 4*, *Rte. 4*, or “Route 4’’” (double quotes from inch marks). I recommend trying these exact-phrase variations in the 1949 Gadsden Times:

* `"John Burton LeRoy"`, `"John B. LeRoy"`, and `"John LeRoy"`
* `"died at his residence"` and `"died at his home"`
* `"Gadsden Route 4"`, `"Rt. 4, Gadsden"`, `"Rte. 4, Gadsden"` (with/without comma)
* Search the **issue dates 12/14/1949 through 12/16/1949**; if no hit, expand a week either side.

If you’d like, give me any subscription screenshots/clippings you have (or a narrowed date window), and I’ll transcribe/verify whether the line appears verbatim. Or, if you prefer a librarian route, I can draft a concise request email to NEAGS specifying the date range and phrasing you want confirmed.

[1]: https://ancestors.familysearch.org/en/LHDR-BP9/john-burton-leroy-1878-1949?utm_source=chatgpt.com "John Burton LeRoy (1878–1949)"

[2]: https://en.wikipedia.org/wiki/The_Gadsden_Times?utm_source=chatgpt.com "The Gadsden Times"

[3]: https://en.wikipedia.org/wiki/Howard_Gardner_Nichols_Memorial_Library?utm_source=chatgpt.com "Howard Gardner Nichols Memorial Library"

I was actually pleasantly surprised with the results. It had been a year or so since I last asked ChatGPT any questions like this. Previous attempts did not provide citations or links. This time, ChatGPT acknowledged it could not find the exact search phrase I entered (and did not appear to make anything up.) But it provided reasonable and solid direction I could take to attempt to locate the information on my own. And most important to me, it provides references, links and citations so I can double check the data presented, and to assist me in trying to find the information I'm after. Both results were helpful, and the slight differences highlight the importance of trying again when you're still staring at a seemingly impenetrable wall. 

While AI is still imperfect, it's starting to become significantly more useful. Over the coming months and years, I fully expect we'll see incredible progress in how it can aid us as researchers, and its integration into websites and programs we already use. I fully realize not everyone is on board with AI. And yes, there are definitely environmental concerns to be considered as larger and larger data centers are required to handle the intense computing demands of AI. But the reality is - AI is here to stay. And in my very limited experiment, it provided useful research direction. In addition to the ways I already us it, I'll continue to check from time to time to discover new advancements in this powerful technology. 

03 September 2025

James Oliver Fowler's Last Sleep

James Oliver Fowler
By now, you know the little angels who passed away too soon tug at my heart. James Oliver Fowler is no exception. His brief existence, and this solitary post-mortem photo, a haunting reminder of just how challenging times were during the depression.

James was born approximately June 1930. I say "approximately" because his father, who was the informant on his death certificate, wasn't sure of his exact date of birth. The family had seven other children at the time (ranging from 1 year old to 17) and, understandably, during the stressful time of losing a child, it isn't surprising he may not have been able to recall the specific date. His birth date is shown as "don't know" and his age "about 4 mos."

1930 was toward the beginning of the Great Depression. It heralded a particularly difficult time in the poor communities of Georgia where James was born. It was an area of tremendous poverty and a scarcity of food and resources. Children (as well as adults) often suffered from infections and malnutrition. Families may have delayed seeking medical care for financial reasons. And even when they did make the call for the doctor, or a trip to a hospital, the level of care was far different than today.

James Oliver's cause of death is listed as marasmus and diarrhea, with malnourishment as a contributory factor. Together, these were a deadly triple whammy. Marasmus was a widespread and often fatal form of malnutrition. It caused extreme wasting, with loss of muscle and body fat. He may have suffered from other infections, compounding the problem. His mother may also have been malnourished and simply wasn't able to provide him the nutrition he needed to thrive, I can't imagine the terrible loss, and most likely guilt, his mother must have felt. I hope, with time, she knew she did all she could for him. 

When we think of a 4 month old, we think of a round chubby cheeked little cherub, with pudgy hands and soft squishy legs. But in this post-mortem photo of James, generously shared by an Ancestry member on is father's side. he looks almost like a fragile porcelain doll, his little fingers curled in the familiar grasp of an infant. His eyes are closed, as if he's peacefully napping. This was known as "the last sleep" pose, and this is likely the only photo ever taken of James. While we're somewhat unaccustomed to this type of photo today, it was not at all unusual for the time. It was a way for the family to immortalize the child. It was a memento to be treasured. I'd like to think a copy of this photo was kept in their home, and he was never forgotten. But it just as easily may have been tucked away, quietly taken out as his parents worked through their grief - a poignant reminder of the son they lost.

James Oliver passed away on October 7th, 1930 at 9:00am. He had been under the care of a physician for almost a week. From the death certificate, it appears he may have been seen at home. There is no hospital listed and RFD is shown in the address (likely the rented farm they lived on.) By the time he was seen, it was unlikely there was anything the doctor could have done. His death certificate says he was buried on October 8th, 1930 in Bethel cemetery. I did find a Bethel United Methodist Church cemetery in Dalton, Whitfield, Georgia. Sadly, there's no mention of a memorial for James. There were 9 other unnamed infants buried there, but none with the surname Fowler. 

I am grateful to the Ancestry user who publicly shared this photo, and for her generosity in allowing me to use it for this post. This is the essence of genealogy, of sharing the details we learn with others. Were it not for her gracious spirit, James Oliver Fowler, born after the enumeration date for the 1930 census, and not mentioned in the obituaries of his parents, might have been remembered only by his brief entry in a death index and his death certificate. This lone photo reminds us of his short, yet precious life.

Sleep in peace little one. You will not be forgotten.

Photo courtesy of Melissa Hunter, graciously shared on Ancestry

01 September 2025

The Photo Hint I Wish I'd Never Saved

Ah yes...the bell icon notifications on Ancestry. 
We have NO way to turn them off. NO option to filter the type of notifications we receive. And most annoyingly, it seems  some hints simply repeat over and over again...ad nauseum.

I dismiss 99.9% of these notifications outright without even reading them. About the only ones I find interesting (or useful) tell me how many people I added and how many records I worked with the previous week. Otherwise, they're almost a complete a waste of time for me. However, the very few I do want to see would be buried if I didn't regularly clear all the others. So, I dismiss over and over...and over again. 

But the worst is a photo hint I saved a couple months ago from someone else's tree. I get notification after notification there may be a new photo in hint in their tree. I dismiss the notification several times a day, only to have it keep coming back. I mark it read before I dismiss it....and you guessed it...it comes back. I've clicked and gone to the tree, but no joy. I've probably dismissed the same notification (shown above), 300+ times in the last two months (and I'm not exaggerating.... upwards of 5 times times per research session the same exact notification pops up, day after day after day.) I truly wish I'd never saved any photos from the specific tree in question. It's not even a tree I'm a guest or contributor to - it's just a random public tree I saved a photo from. It's almost like this one notification is stuck in the queue. I've saved many other photos and they don't haunt me like this one does.

Please Ancestry - give us an option to disable or at least filter these notifications. Or even an option to permanently delete or mute specific ones would help. Make it a ProTools feature. I'd pay for ProTools forever if it meant I could get rid of these notifications (or at least stop this one notification from haunting me day and night!)

In general, I love Ancestry. I really do! But this feature is not a help for me and it's not a time saver. Quite the opposite. I fully acknowledge, maybe it's just me and how I work. Maybe others DO find these notifications helpful. I wish I could just ignore them completely. Unfortunately, it's not how my brain works. I need a tidy screen. I have the same challenge on my cell phone. I can't stand badge counts and have to clear them straight away. It's most definitely an OCD thing....but it's a real thing, even if it's not something everyone understands. You never know, with the ability to customize these notifications, I might even embrace them.

Maybe I'm wrong. Maybe other users truly do find the feature useful. Or maybe, many just ignore the icon completely. I just wish Ancestry would give us options . And now, with my frustrations vented, I shall jump down off my little soapbox, tuck it under my arm and scurry back to my research....right after I dismiss he same photo hint yet again...