Me and My ChatGPT: Living in the Age of AI

“The first reality—and perhaps the most worrying—is that we are using tools that are fundamentally beyond human comprehension.

We have a name for this: the interpretability problem. Nobody—including the scientists and philosophers at OpenAI, Google, and Anthropic—can precisely explain why a massive neural network chooses the specific combination of weighted parameters that results in an AI’s response to a prompt.[1]”

AI is the thing that finally pushed all thoughts of the Epstein files out of everyone’s minds and even relegated a shocking restriction of the free press by the president to the backwater of the news cycle for a moment, so I suppose I should step up and confess it: I use it but I don’t understand it.

Which is like most of us probably. Almost everyone has used it too, maybe not even realizing how much is powered by AI these days. Even if you don’t want to, you sometimes are boxed into it, like my cable company’s less-than-stellar chatbot that is just about the only way to reach them. If you’ve ever called out to Siri or “Hey, Google,” you’re accessing AI. It’s been a part of our lives for a while now.

I didn’t use it to actually write the book or these posts (as you likely can tell since they’re often not that polished), but I admit I did absolutely use AI powered tools to help with research, and with how to draft the book proposal. It also did the heavy lifting on the website design. All things I might have had to pay people to help with in the past was at my fingertips and free (well, the web design came with the cost of hosting the site). For a first-time author with no budget, it was heaven sent. Google Gemini could take my search requests and summarize the applicable sites and then list them so I could refer to them and read the articles. I found many of the books I read using Gemini. It was the research assistant that I could never have afforded, and it was – with very few exceptions – spot on accurate. When I turned to ChatGPT for help on how to craft a book proposal, I was also struck by how eerily intuitive, accurate and, dare I say it, conversational it was. Like texting with a person whom you know and who likes you, but is way, way smarter than you are.

That was different than my first encounters with it. Going way back in time to, say, two years ago, when our department heads were wanting to leverage AI for several repetitive tasks that were slowing us down and expensive and prone to error since they relied on humans, we were all encouraged to learn it and use it. Of course, the other, smarter and younger people on my team dove right in and were using Claude and some of the AI options that were more tech-oriented to help with writing code, but that was well beyond me, so I tried out ChatGPT to dip my toe into the water. The first thing I asked it was to create a picture of Godzilla smelling a flower.

I was pleasantly surprised with the result. But when I tried to up the game and ask it more complicated queries to help with training materials and presentations, it often came back with questionable results. Formatting issues, misunderstanding the mission, and sometimes downright gibberish were all common issues. My boss said it was all in the quality of the initial inquiry, but using the free version of it, you’re limited to how many times you can prompt it in a day to try and refine that. I did watch him, however, try multiple times to polish queries on various AI tools to narrow down and get the results he wanted. Like a kid in a candy store, it excited him. To me, it seemed like a lot of work. But all that work was serving to train the tools, and they were eager learners, and now, ChatGPT, even the free version, is so much more sophisticated that, while I was extraordinarily grateful for the help, it did give me a bit of a pause. I recognized that this was nothing like my initial foray into the same tool, and it was daunting to think about how fast it had evolved. Particularly since the chatbot my cable company uses still completely sucks, just like it always has. Therefore, when the dire warnings broke directly from the people who fathered AI, I cannot tell you I was surprised. And I believe them.

I had also just finished reading American Prometheus, perhaps ironically, which details the life and times of J Robert Oppenheimer, including his campaign to caution Washington on the proliferation of nuclear weapons in the Cold War era, and it all seemed prescient. I saw a lot of parallels to the warnings being given now by Sam Altman and Dario Amodei. Both are smart men; I’m sure they’re aware of Oppenheimer’s saga, so their willingness and sense of urgency to put some guardrails on AI, should be taken seriously by everyone. They’re willingly putting themselves on the same potential path that Oppenheimer traveled, risking being made pariahs by those who want the pace of AI learning to continue, greedy to cash in on its potential uses. Some might still say that Oppenheimer was the boy crying wolf because we have not destroyed mankind with an H-bomb or the rest of our considerable nuclear arsenal. But we could, and his concerns and warnings are as valuable now as they were when he was making them. Perhaps my father’s bosses in the Air Force’s Strategic Air Command, who were on the opposite side of his arguments, did too, but he had the right to voice his opinions in a free society, and he felt that he had a responsibility. Trying to silence him and distance him from decision makers in Washington is a black mark on our history as a democracy.

As was true in the early nuclear age, most of the public doesn’t truly understand the topic they’re hearing warnings about. I certainly don’t. We know that it uses extraordinary amounts of energy, hence the advent of the controversial data centers. And we’re being told it could destroy mankind, but that is such a grandiose statement for an average person to hear that it is easy to dismiss. I’m probably not the only one who thought of Robert Oppenheimer and how we have managed to exist in a nuclear age to-date. But now we’re introducing artificial intelligence we don’t fully understand into a nuclear age. If we lose control of the one, it could unleash the other. I think I will choose to believe the men who ushered AI into our lives when they talk about it and urge caution.

The most important thing to know about AI, I've concluded, is that our elected representatives will put appropriate guardrails in place to keep it in check. We all need to think about it when we go to the polls this fall, and we need to vote for politicians who are willing to learn and believe the warnings that are coming out like our lives depend upon it, because they very well might.

[1] Scheinfeldt, Tom. 2025, November 12. “Ten Things You Need to Know About AI.” FoundHistory.org. https://foundhistory.org/ten-things-you-need-to-know-about-ai