FamilySearch.org
During the past weeks, I have fielded a number of inquiries at the BYU Library Family History Center about the reliability and viability of the FamilySearch.org AI Research Assistant. It appears that there are two iterations of the AI Research Assistant with the one above that appears on the startup page and another one that appears on individual profile pages.
The startup page AI Research Assistant finds suggested areas of research, while the second one looks at finding records for the profile page person specifically.
I have tried to use the startup page AI a number of times. In the image above, I clicked on the entry for the Possible child of Frank James Hamilton 1847 -1932 L5NX-Q86. Clicking on the entry brings up the following statement.
First of all, at this point, I know absolutely nothing previously about Frank James Hamilton. I have no idea how I am even related. The name seems to have been picked randomly from the FamilySearch Family Tree. I opened another tab and looked for Frank with his ID number.
I haven't done any work on this line for some time but I knew this line had some problems. Starting with Eliza Ann Hamilton LDWV-RK4 it appears that the only record showing her father is a marriage record in Nicholas County, Kentucky showing her father as "Jas Hamilton."
So far, the AI suggestion is a long way off from being my relative. The real basis of the concern however, is that some users of the Family Tree would look at the suggested entry and be ignorant or not take the time to examine the relationship. I guess this doesn't matter if you don't care who you are researching.
Now, I could keep going and see if there is a valid parent/child relationship for James P. Hamilton LZKR-2QJ but that is rather simple, there are no such records currently attached. So where did his Family Tree father, John Hamilton 27H5-KQ7 come from? John has 4 sources attached; two deeds, a will, and a Sons of the American Revolution membership application. The deeds are from Elizabeth Hamilton and do list a children named James and Jason. I uploaded the deed back in 2022. The other reference to a deed is only a reference without a copy of the deed attached or further identified.
Now, we need to look at the reference to Jason Hamilton who is supposed to be the father of the AI identified Frank James Hamilton through William Hamilton. Jason Hamilton LBRL-SW4 has three sources, Kentucky Marriages, and no sources showing any children. The next person in line, is William Hamilton L5N7-QNV, who is supposedly born in Kentucky and dies in Illinois. The only connection between this William Hamilton L5N7-QNV is a census record stating he was born in Kentucky. The 1850 Census attached does not list a Frank James Hamilton but does have a James Hamilton who was born in Illinois in about 1845. Was the James Hamilton born in Illinois the same as the AI suggested Frank James Hamilton? There is one problem, none of the records attached to Frank James Hamilton show his name as Frank James Hamilton.
Well, here is main issue I see with the AI Research Assistant's suggestions, they rely on the accuracy of the records already in the FamilySearch Family Tree as being correct. Do you really want this kind of suggestion and will you take the time to examine the entire ancestral line in the Family Tree before adding the suggested person to "your family"?
Good Luck.
No comments:
Post a Comment