Flagging transcription error from FMP record
Flagging this because it's a doozy, though I'm not sure what can actually be done about it as the original isn't held by FS. The error looks like it's lifted straight from FMP, who also has the same transcription error, though in more detail.
This record for the baptism of a James Burgwin in Radnorshire, Wales (30th May 1735) is essentially a fabrication. No such baptism occurred.
The image of the original record, when viewed via the link to FMP, is actually an entry for the marriage of "John James" and "Anne Burgwin" on the 30th May 1735. (FMP list "James" and "Anne" as parents of "John")
It's also listed as appearing on page 21, though that is the facing page. (I checked both pages for the alleged baptism, it doesn't appear on either).
I don't know if AI is responsible for the transcription error since it is so wildly inaccurate, but I suspect it may have something to do with the fact that the original record set appears to be classified as "Radnorshire Baptisms" when in fact the original ledger contains baptisms, marriages and burials all mixed together.
What concerns me is, if this isn't an isolated incident in this record set, there are potentially numinous phantom baptisms polluting the available data which could lead to some pretty significant errors across both the tree on FS and in peoples personal research, particularly as the original isn't readily available to view on FS.
As I say, not sure what can be done, but flying it up the flag pole for those that do.
Answers
-
@RaniM - you might have already twigged this, but for other people's benefit, it's fairly standard for FMP collections to be named as "XXXshire Baptisms" (say) but for the original paper registers to be multi-event. Which does lead to potential issues if someone slaps a baptism template on a page containing marriages and / or burials.
My gut feeling is that this isn't an AI related issue as FMP did the Welsh registers years ago - which doesn't mean that additional registers might not have been done since.
It's definitely an odd error, though, and makes me think about an oddity I found in the Betley (Staffs) indexes on FMP. This was one of my family names, whose baptism was allegedly on a certain date. There is no sign of the original for that index on FMP. I checked the PRs, the BTs, even a Victorian transcript elsewhere - no sign. But when I was checking, I only looked at the baptisms, I never checked the burials or marriages. Hmm 🤔
3 -
In my experience, FMP is responsive to reported errors. For example, a couple of years ago, I encountered a County Sligo, Ireland, marriage register where the surnames had been misaligned. It is a better-than-usual register because the full names of all 4 parents are included. Unfortunately, the birth surnames of the mothers had been attributed to the contracting parties.
After a couple of email exchanges, they agreed that the problem needed repair, and they did fix it.
I raised the issue by email to support@findmypast.co.uk, providing several specific examples.
For a one-off correction, the process is even simpler, using the "Report an Error" button, hidden under the "More Actions" button.
The FS versions of many collections resident on other sites are refreshed periodically. The collection for the Wales, Radnorshire registers was last updated on 8 December 2025.
4 -
Thank you for your insights @Adrian Bruce1 and @Áine.ní.Donnghaile, I can always count on you to come through with the knowledge!
I've emailed FMP notifying them of the problem and hopefully they'll be able address the issue, thanks for the email address @Áine.ní.Donnghaile. I tried the "Report Error" option, but that really only allows for corrections within the prescribed template, nothing to report the template itself being wholly incorrect!
Fingers crossed that will then trickle down to FS, once amended.
@Adrian Bruce1 I tend to agree with your gut re: AI, but it is just such an egregious error I couldn't see a human making it. However, your thought about template use makes a lot of sense and if an early image-to-text program was used in conjunction without oversight, I can see this occurring.
I do wonder if your mystery baptism in Betley was a similar phantom. I suspect it would be hard to know without manually going through the original images around that time, particularly if the original names are being spliced into whole new individuals in the index.
2 -
@RaniM If you're comfortable, please let us know how it goes with FMP Support.
When I reported the County Sligo register issue, I received a first reply within 24 hours. It took a few months for them to re-index the register.
1 -
@Áine.ní.Donnghaile more than happy to let you know.
I've already heard back from FMP Support saying they've notified the data team of the error, and once it's been fixed they will let me know. I was impressed by their promptness.
It will be interesting to see how long the issue itself takes to be addressed, in light of your experience. I'll update when they get back to me again.
2 -
Thanks @RaniM
For reference, I reported on 3 August 2023 and received the first email response on 4 August. I received a second follow-up email later that same day.They had said "we will keep you notified of any updates," but I didn't hear again until this announcement in the Friday Blog of 8 December 2023
we've also improved one of our existing Roman Catholic collections. Containing over 3 million records, this set is a key resource for anyone with Irish Catholic relatives. We made significant improvements to 19,000 records within this collection - improving their transcripts, and identifying names that were previously missed out. This means that there are more names than ever to search for, and your ancestors may just appear where they didn't before.0 -
Bit cheeky of them to promise updates and then… not! Haha. That's good to know though. The first response they sent me was within 12 hours of my email, which I thought was excellent. But in light of your experience, I might follow up with them myself in a week or so, if I haven't heard anything. I'll keep you posted!
1 -
Little update @Áine.ní.Donnghaile and anyone else following along:
I've just received another email from FMP confirming the issue is valid, an investigation is underway, and that it appears to be affecting multiple record sets within those baptisms (as I feared). So the issue may take longer than usual to fix.
So if, in future, the FS Community starts getting questions about missing Welsh baptisms that suddenly can't be found, but records previously existed for, this is why. It's not a licensing issue, they weren't real baptisms to begin with. It's also worth noting that it's very possible PIDs have been created from these false records.
It does make me wonder what other issues are out there that haven't been found yet. It just proves the point that you should always, always, view the original before taking something as fact.
In light of that, I don't expect to hear again for a little while. But to see the exercise through, I'll update again once I do.
1 -
Thanks for the update, @RaniM
1
