Indexed Link glitch for DGS 007719893
Answers
-
I agree there is an error. I Looked at one record from the DGS quoted above https://www.familysearch.org/ark:/61903/1:1:66KV-2DBF?lang=en which gives the christening date Event Date 27 de julio de 1908, yet I looked at the image I could see 1687 at the top of the page. The image was covered with yellow, perhaps this is a record which has been indexed by AI?
@Ashlee C. could you investigate/pass on?
https://www.familysearch.org/en/search/catalog/results?q.filmNumber=007719893 clicks through to https://www.familysearch.org/en/search/catalog/21321
then you can click through to index records https://www.familysearch.org/en/search/record/results?q.filmNumber=7719893
0 -
@LR Heyn I have been assured that AI indexing will get better and better, but for now it's going to be all hands on deck to fix the mess ups. The date errors on this DGS is not as bad as it could be—about 5 percent of the 12,000 christenings, which is not out of the range for users to edit. I think the 1908 came from a word following the month that looks like "ocho".
The bigger problem is mistakes in identifying and transcribing correctly the name of the principal and relationships. Looking at the example provided by @MaureenE123 right away you notice there are two fathers and no mother.
Here is the image from the film:
Child is named Juan Christoval Núñez (not Christoval Alvan), father is Matheo Núñez, mother is Maria Nicolasa, year is 1687. I was easily able to edit the child's name and the date, but apparently we are still having problems with relationship editing. I was able to remove and replace the parent (mother) but it would not save and reverted back to the incorrect father. I have been working with engineers on these glitches and thought they were fixed. I will report this one.
1 -
Given that the example detailed above has multiple errors, and you say
I have been assured that AI indexing will get better and better, but for now it's going to be all hands on deck to fix the mess ups.
I believe that the records which have been indexed by A.I. should advise this on all applicable records, and indeed add the proviso
"Having been indexed by A. I, this record may contain inaccuracies. If you find any inaccuracies, please edit the data."
I strongly feel this is necessary. For countless years people have known the indexing went through a two stage process, with the original indexer and then a reviewer, so there was a level of trust in the results, in that the record had been seen by at least two people.
Now it appears that A.I. produces a record where there are absolutely no cross checks at all. Eg in the record set above the baptisms were for the period 1682-1763, yet there is no checking that picked up results were showing events 150-200 years outside this range.
Even the name was incorrect in the example above.
Unless there is something added about the A I Indexing, at the very least there will be confusion about the situation, as detailed in the original post above, but I think there is a possibility that people will lose trust in the reliability of the FamilySearch database, and stop using it.
1 -
@MaureenE123 Well said. I believe calling attention to the errors will make users more vigilant in finding and correcting them.
0
