More Reflections on Doing Genealogical Research with AI Assistance

 

Comment on this AI generated image: I realize that a substantial number of people use their laptop as their main computer, but many of us also have more than one monitor, a tangle of cables to various hard drives and peripherals, and other random items on our work desks including, but not limited to headphones, mics, random papers, cough drops, and boxes of tissues. So, this picture is far from reality. 

A comment to my last post on Facebook got me thinking. Progress with any type of research depends heavily on your own ability to do the research. When I first started looking for answers to my questions as a nine- or ten-year-old child, my "research" was limited to the card catalog at the Phoenix, Arizona public library. I soon learned not only how to find what I was interested in while using the card catalog, but what was missing from the library's collections. My involvement with research and libraries continued with my university experience as a Bibliographer at the J. Willard Marriott Library on the campus of the University of Utah. Both my studies and my job required constant research. My initial experience with computers began with learning how to code a Shoshone/English, English/Shoshone dictionary onto the mainframe computer also on campus. Here I am many years later still doing exactly what I learned to do starting as a young child. 

By the time I started my current genealogical research, now 44 years ago, I was already doing legal research as a trial attorney and had access to the Phoenix law libraries, the library at Arizona State University, and my own collection of books. I also, immediately upon their introduction, began using the various home computers beginning with a TRS 80. As the internet came along, it merely expanded my research interests. 

Fundamentally, to do research on a computer connected to the internet, you still need to know how (and also why) to do basic research. When all this AI started showing up online, I immediately realized that you not only have to know how to interact with the AI Chatbot (Gemini) but you have to understand the responses. Research is research whether you are seeking information from a book somewhere in a large library or from a chatbot on the internet. To put the concept of research into the most simple terms available, it is this: you have to ask a question and then go find the answer. Now with the internet and with some assistance from various chatbots, I will have more questions and even more answers.

Now a comment on the technology. I will use whatever technology appears to be the best at answering my questions. Today, that appears to be Google's Gemini and NotebookLM combination. But that may change tomorrow or even later on today. I use the same tests I used in the libraries long ago. I looked for something I knew existed to test whether the library or whatever had the information. 

Today, to test the AI responses, I use examples that I have been using for the past 44 years; my Great-grandfather Henry Martin Tanner and my most remote Tanner relative, William Tanner. In both these cases, I do some preliminary research using the Full-text search capabilities of FamilySearch for finding the initial "new" to me documents. I already know the questions to ask and what to expect if anything works. I have been using both these test subjects for years for evaluating search engines like Google and web browsers such as Chrome, Safari, FireFox and etc. The key here is that AI research is sort of like working with a mirror of your own research abilities. You only get back what you can add and ask. Otherwise, it is the same things I would normally be doing for research that I have learned over my lifetime. I realize this isn't a very comfortable response, but you do have to come to the table with your own skills before you can make much progress. AI mainly accelerates the process and gives you more to work with than you ever believed possible. But, it doesn't answer all the questions. You have to do the research work first and last. 

Finally, for this post at least, you need to realize the vast amount of knowledge that the internet and the computers do not have "digitized" for consumption. This includes the vast amount of information in libraries and archives that is yet to be digitized. Watching the growth of online research opportunities is like watching a giant backhoe tear down a building, there is a lot of noise and it is fun to watch, but the real work doesn't begin until the new building is designed and built and meanwhile you have a lot of garbage to plow through to get to the answers.