For tonight's Saturday Night Genealogy Fun challenge, Randy Seaver has posed a question that has actually been occupying my brain for a while.
Come on, everybody, join in and accept the mission and execute it with precision.
1. Are you experimenting with artificial intelligence (AI) for genealogy and family history? What have you learned so far? What have you done to date? What GPTs have you used? What results have you had — good or bad?
2. Tell us about your AI experience in a comment on this post or in a Facebook post. Please leave a link on this post if you write your own post.
I'll start out by saying that I am not going out of my way to experiment with artificial intelligence. I have used no GPT's or set up any searches, no matter how much Bing keeps pressuring me to do so. Google was bugging me about it also but seems to have given up. I signed up for Bard at the beginning but decided it wasn't for me.
I know that I have been experimenting with AI anyway, whether I want to or not. It's built into the hints that the genealogical database sites use, it's permeating our search engines and their algorithms, and it's being integrated throughout the computer and business worlds, because it's the latest greatest coolest thing that's going to make our lives easier and more wonderful and solve all the world's problems
Oh, wait, that isn't true? Gee, you'd never know it from the way it's being touted.
Color me a skeptic, but I read an insightful article recently (and I'm still trying to figure out where I put it on my desk, which is why I haven't already posted my thoughts about AI) which pointed out that not only, as Randy mentioned, does AI give incorrect information when it doesn't know the answer because it just has to tell you something, but relying on AI to do your thinking for you means you are not using your own brain to do your thinking. Go down that path often enough, and how long will it take for your brain to need AI for all those answers?
This is the stuff that many science fiction stories are based on: technology being oh so convenient and performing all those boring tasks for people and people becoming more and more reliant on technology, to the point where it's no longer a choice to rely on technology, because people are no longer capable of doing those tasks themselves.
Using Randy's examples:
• Writing a biographical sketch: Ancestry can already do this when you post a tree and input some dates (I'm sure that's classified as AI). I'm pretty sure other sites and databases can do something similar. Or, heaven forfend, you can write one yourself, using the same data. AI will write in a standard formulaic manner. You can actually give it personality.
• Writing an obituary (yours or someone else's): Yes, you can have the computer create an obituary for you, or (again) you can write it yourself, using the same information.
• Book recommendations: You can do this with a search engine already. Search for books on a given subject, and you will receive a list longer than you want to look through. You can do the same thing on Amazon. I admit this is a situation where it's helpful to have that kind of capability, because you can't find what you don't know about. But this capability has existed for years and years. We didn't need ChatGPT to do it.
• MyHeritage features: So, as I wrote above when commenting on Ancestry, another site can also create a biographical sketch based on information that has been fed into it. RecordFinder sounds suspiciously like twitching leaves on Ancestry and hints on FamilySearch.
• Ancestry's AI feature: Wow, I can input a search like that into Google and get something similar. And Google is drawing from many more sources than Ancestry. Actually, Google searches Ancestry also.
• DeepStory: I already find it frighteningly misleading the directions that MyHeritage is taking on having moving and talking photographs and use of multimedia, because it won't take long for people to forget that they're all computer-generated and are not our relatives, just someone's interpretation of what they might have been like. I want reality, not someone else's version of what it possibly could have been.
• FamilySearch's full text search: I find it ironic how this is now being received, when not that long ago at an early RootsTech — you know, that genealogy conference that actually used to be about tech in genealogy?; nowadays referencing one URL in your presentation counts as tech — an app for reading handwriting came in second(!) in the innovator challenge to something as banal as being able to record a phone call on your mobile phone. I guess the church finally decided that reading handwriting was worth something. I have often wondered what happened with the programmer(s) working on this question years ago.
That said, I will admit that it's a lot harder to learn how to read old handwriting than how to write something relatively straightforward such as an obituary or a quick historical profile of a person. But this will quickly become something where people don't check the writing themselves to verify the reading made by AI, they'll just accept it and move on. Gee, I hope it's accurate.
• Google Bard/Gemini and MS CoPilot: You can create your own images/collages with standard photo editing software that already exists. Some of the programs are even free. The more you do it, the better you become. Why let someone else's AI improve, instead of improving yourself?
Another thing to keep in mind is that, along with only being able to provide answers based on what information it has available, AI has already been found to carry on and put forward the preconceptions of the people who program it. So if it doesn't have more than one perspective available, or if one perspective is weighted more strongly, that will affect the answers it provides.
So, cynic that I am, I am not jumping on the AI/ChatGPT bandwagon and have no plans to do so. I know I can't avoid it entirely, and I'm not trying to, but I'll use it judiciously and continue to rely primarily on my brain and my own abilities to puzzle things out, write obituaries for family members when the need arises, make collages, and research what happened in the times and places my family members lived. I'll even continue to write my own blog posts, like this one. After all, how do we know that Randy didn't just feed ChatGPT the topics he wanted to cover and let it write his column?