Family History and Tools of Artificial Intelligence

It is kind of stunning how the term “AI” has taken over RootsTech this year.

AI stands for Artificial Intelligence and it is revolutionizing the Internet and nearly every industry that uses it.

In my book, and I’m going to go on an old man rant here, AI is what gives us toilets that flush just as we are sitting down and maps that take us out to the middle of nowhere and dump us off a cliff.

In my experience, the budding technology of AI is long on artificial and short on intelligence.

In my real-world work life so far AI has proven to be a jobs-killer. In what I do, words are paramount. From crafting a written article to designing a well-written product description, my writer’s brain is all about clear short-form communication that is friendly to search engines and people alike.

AI is taking the keyboard away from people like me.

Imagine it: you write a well-researched article for a website or a publication that takes time and considerable research. It gets posted. Then some AI-enabled bot comes along and scrapes your content, re-arranges a few words, and publishes your stuff without credit or compensation to you. That’s the real world of AI right now.

It is simple thievery at best and plagiarism at worst. Artists of every stripe are affected: writers, musicians, sculptors and, yes, even photographers (and many others).

How does AI fit into family history?

That is a very fair question.

You see, there are parts to doing family history that I do not care for – specifically, the heavy lifting of data. You know, the stuff of names, places, and dates we attach to every family member to document their lives.

The genealogical part of family research is tedious. It requires a detective’s mind and strict standards of sourcing. There are books, classes and even college degrees about this stuff.

The reason I don’t care for this part of family history is that I lack the training, the knowledge, and especially the discipline to do it right. Like doctors, it takes someone special to do this work well.

FamilySearch, which is the clear center of the family history world, is embracing AI to make the tedious work of genealogical data easier and faster for family historians.

This is a good thing if you think about it. If FamilySearch can use AI to generate better search results why wouldn’t we embrace that?

They are using AI to read old handwritten records to index them without human eyes. That’s okay, right? Sure. As long as it is accurate and does not waste time for someone digging for the truth.

At the end of the day that kind of use of AI is useful and doesn’t compromise the needed accuracy of traditional genealogical standards.

But AI at RootsTech in 2024 is going way beyond these obvious uses. Like everything AI everywhere, RootsTech is infested with AI gone silly.

How silly?

There are vendors hawking AI-generated chat bots that emulate your ancestors. You can actually talk to your ancestors via AI.

Do you feel comfortable with that?

There are now AI-driven programs to write your family history.

Do you see where this is going?

Let me give you a visual of how AI is being used right now and why is it not necessarily a good thing for family history.

Here is an image we all know of Edwin Ruthvin Westover. It’s not even a photograph. It is, ironically, an image made by hand many years ago by a family member interpreting what Edwin might have looked like. Like many, I have appreciated this image:

Well, we already know this isn’t really Edwin. It’s a drawing. So what could possibly be wrong with using another interpretation? Check this out:

Using this image, plus a few words that sends the AI-generator out on to the Internet, it returned this new photo-realistic image of Edwin:

Is this a good thing?

Imagine I put this on Ancestry or on Family Search claiming this is Edwin. All it takes is one person to do this and it can never be retracted.

We see it over and over again. In fact, I sometimes get myself in trouble by identifying images that are inaccurately posted. I once had the nerve to declare an image false because it was posted as being an individual who was born in 1715.

I didn’t have to work hard to prove it – the first actual images of humans didn’t surface until the very late 1830s.

Despite pointing it out that image continues out there online attributed to the person it can never actually be, attached to who knows how many family trees.

What happens when AI using a real photo instead of a drawing? Well, I ran this well-known and proven image of William Rowe:

Here’s the AI William Rowe:

I don’t have a problem with AI used as a tool to help me find true information.

But I have to draw the line at using AI for anything other than a tool.

Truth in digging through the past is challenging enough to find.

Generating mistruth muddies the waters way too much for family history.