Artificial Intelligence – or “AI” – is all the rage these days. If you use a computer, a smartphone, your car’s navigation system, or have called a company for technical assistance, you have used AI. It is becoming ubiquitous in today’s society.
So what exactly is Artificial Intelligence? Nothing new, really. But it has been evolving at a rapid pace. In 1950, the British mathematician and computer scientist Alan Turing proposed a test to determine if a computer was “intelligent.” Called the “Turing Test,” it basically proposed that if a human being could have a blind conversation with a computer — and was not able to tell if he was conversing with a computer or another human, then the computer was intelligent.
Modern computers began passing this test around 2010. And today’s supercomputers can easily fool anyone with their perfectly phrased and educated responses. In fact, it could be said that the latest AI models should flunk this test because they know so much and answer so quickly, that no human being could possibly be at the other end of the conversation.
While AI technology has been with us for over a decade now, acting as a navigation or home assistant, performing facial recognition, helping with internet searches, and even beating grandmaster chess champions — it has only been since late 2022 when we could chat conversationally or brainstorm with an AI client. For those who desire to do so, AI platforms like ChatGPT, Google Gemini, Claude, Microsoft Copilot, or Meta AI open up vast resources for researchers and everyday people looking for answers to difficult questions.
Most people are amazed at the capabilities of AI. Not only by its humanlike responses, but its depth of knowledge about every conceivable thing you can think of. This is accomplished with a technology known as “Generative AI,” which uses massive neural networks modeled after the human brain. Huge data centers are fed massive amounts of data, then use machine learning and comparison to generate a string of words based on probability to fit the given context of the question.
The benefits of AI
There is little doubt that AI is a valuable tool that can be used for many purposes. Most people who choose to take advantage of the capabilities of Gemini or ChatGPT will use it to get answers to everyday questions, help plan a trip, or find the name of an actor in a favorite TV show.
My first experience with AI was asking it for help in troubleshooting a home thermostat a couple of years ago. The ability of Google Gemini to engage in a conversation where prior inquiries were remembered and augmented was amazing to me. Once I had told it the make and model of the thermostat, I could then ask it things like “where is the yellow wire supposed to connect?” and “what does the LED on the circuit board mean when it is flashing” — without having to tell it the device model number again.
Since I have a background as an IT professional and programmer, I find AI particularly useful in troubleshooting software and website problems. Have a problem with a Windows error message? Just ask Gemini for a quick answer. “Wondering why the picture on your webpage isn’t the size you want it to be? Just feed the HTML source code for the entire page to Gemini, and it can tell you the problem. Pretty amazing, really.
As an author and historian, I also find AI to be helpful. If I need inspiration for an article, Google’s Gemini (my go-to AI client) can give me some ideas. When I wrote my article “Boom, Bust, and Vanished: The Town of Fivay,” that was previously published in the Suncoast News, I was having a hard time coming up with a title. So I fed the entire 1,000-word article to Gemini and asked it for title suggestions. It came back with nine suggestions to choose from, categorized by genre — did I want a title for an article in a history journal, a newspaper, or something sensationalized for a tabloid paper?
I’ve found two ways where AI is useful in my historical research. It can be helpful in developing an outline for an article. Gemini has a useful way of organizing things into bullet lists. And they can be helpful in figuring out how to logically describe a particular subject. The other way I use Gemini comes at or near the end of my research. Since AI is really good at finding essentially everything out there on the internet about the subject at hand, I use it to see if there is anything I am missing, or that I have not yet discovered. The stuff it finds must then be thoroughly fact-checked against known sources — like newspapers, books, firsthand accounts, and other traditional historic archives.
The dangers of AI
Now that you have my personal opinions about the benefits of AI, it’s time to explore some of the potential problems AI poses. Many people feel threatened by this amazing technology that has sprung up so fast — and rightfully so.
Most of the recent negative news reports about the influence of AI have to do with the public fear and resentment of the huge data centers needed to power the technology. About the amount of water and electric power they consume, and the ecological effect they may have. There is also concern about the AI revolution eliminating jobs held by humans.
My main concern is about the effect of the technology itself. The ability of newer AI models to produce pictures and textual content that is virtually indiscernible from real photos and human written articles is troubling. This is particularly worrying when it comes to things like “deep-fakes,” political advertisements, and social media posts. It is now fairly common for “authors” to produce pulp fiction books that are completely written by AI. And it is perfectly legal to do so. Using AI-generated content, and calling it your own, is not generally considered to be plagiarism since you are not stealing it from another person.
My biggest concern, from a historian’s perspective, is that AI will begin to erode the factual authenticity of our collective historic knowledge. People can now ask Gemini, or their favorite AI client, to write a 500-page article about some historic topic, and then post it on social media without any follow-up fact-checking. It then becomes part of our collective knowledge which is out there forever.
And don’t be fooled … AI very commonly lies to you when you ask it a question. Fact-checking and common sense must be applied to everything an AI client tells you. As an example, I asked Google’s Gemini about using it for historic research, and this was its response:
“History is full of gaps, dead ends, and people who left no trace. Artificial Intelligence (AI), by its nature, hates a vacuum. If you ask it about a gap in a genealogy or a historical event, it will “bridge” that gap with a plausible fiction. Over time, the “bridge” becomes the road, and the original “gap” is forgotten.”
This is like the “Ministry of Truth” from Orwell’s novel “1984”. But it isn’t a dictator changing the records; it’s a “well-intentioned” AI algorithm trying to find the most popular answer to your question. An answer that will then be propagated as truth by future generations.
It’s scary stuff … but now that the AI revolution is upon us, this genie can’t be put back in the bottle.
Paul Herman is the Digital Media Archivist for the West Pasco Historical Society.