EUREKA!
—attributed to Archimedes
Recently, I wrote about how it seems impossible to go any length of time these days without hearing some story about AI, and the most recent past is no exception. Indeed, the morning I started writing this, NPR carried a story about Elon Musk’s lawsuit against OpenAI’s Sam Altman, and when I googled the topic out of idle curiosity, my screen exploded with other stories. All of which is preamble to say that yet another reason has come to my attention to proceed extremely cautiously with AI’s employment in the classroom. That we must do some deliberate training about it is now as inevitable as dealing with the ubiquity of smartphone technology in students’ daily lives. However, the latest findings about AI’s impact in the workforce—what is being called “AI Brain Fry”—should probably inform that training with more skepticism than some in education are presently embracing.
Why? Because what is coming abundantly clear as AI has been embraced in a variety of work environments is that using it to enhance or augment employee productivity has actually backfired by generating a need for multi-tasking on steroids. As technology columnist, John Herman, of New York Magazine describes it:
You have a programming tool that can kind of run in the background and starts adding features to software really quickly, you have another tool that’s constructing a report from you, it’s searching the web and pulling together a market research document. You have another tool in the background that you’re in a constant chat with trying to refine some idea for a talk you have to give…you’re just kind of getting first pulled in all these different directions, and then you’re kind of spamming yourself [as] you’re just producing all of this product. And it’s harder, as you use more and more tools to keep track of whether this output is actually relevant to your job, whether you’re doing anything that you need to be doing or whether you’re kind of creating new work for yourself…. (Luse, et al)
You get the point: it is exhausting simply reading about everything AI is now asking people in the workplace to manage and keep track of, let alone actually doing it. Again, it is multi-tasking on steroids, and since the brain research on that concept is 100% clear—multi-tasking is neurologically impossible—the whole point of employing AI to improve output and efficiency would appear self-defeating. It’s as if “AI is a poorly trained intern that you have to check the work of all the time, turning workers into bosses or at least simulated bosses” (Luse, et al) who now have two jobs: their original one and their now supervisory one. Talk about a recipe for exhaustion, burnout, and counter-productivity!
Therefore, as we approach the reality of AI in the classroom specifically and AI in education in general, we might want to listen to this early cautionary tale from the frontiers of AI in the workplace. If we do, I think it will lead us to recognize the need for two things. First, what boundaries do we need to place on already common uses of AI such as brainstorming and outline drafting for students or grading and lesson planning for teachers, and second, what counts as true augmentation versus one-more-tab-open-on-the-screen? The brain research on the answer to the first question is so overwhelming—with the new multi-tasking demands from the workplace simply piling on the confirmation—that I’ll simply state that if you do happen to be a first time reader who wants the details, here’s the link.
As for the second question, that’s where some of the more recent brain research gets interesting; so let’s dive in.
We need to start by asking what would we potentially be augmenting with AI, and I would argue that the most logical choice is problem solving. However, we have recently discovered that the human brain has two distinct circuits it uses for problem solving: insight (the “aha! moment”) and logic (“analytical reasoning”), and we have found that each of these circuits has its own unique starting location during resting brainwave activity (insight=left temporal lobe; logic=right frontal lobe). Furthermore, each of us apparently has a hard-wired tendency for which circuit we default to when solving problems (though everyone, importantly, can do both), and in fact, “a few minutes of EEG [readings on a test subject] predicted, up to seven weeks in advance, whether a person would solve puzzles mostly insightfully or analytically. Our predominant thinking style is stable over time” (Kounios & Kounios, p. 24).
Because it is stable, though, the answer to whether a specific variant of AI might augment our problem-solving capacity or interfere with it has big implications for how we have people use AI in their learning. Those who employ logic as their default might benefit from a tool that can look at billions of data points simultaneously to identify the most pertinent ones to employ analytically (e.g. determining all possible gas efficient routes to deliver a collection of packages might free someone up to look at these routes through the lens of that specific day’s traffic). But for those who employ insight as their default, the digital pollution already clogging so many of our inner lives already inhibits the brain’s insight circuits, and any AI augmentation is simply going to overload those same circuits further. AI, with its “unrelenting demand for productivity and speed, denies insight the time and opportunity to work wonders at its own pace” (Kounios & Kounios, p. 27). Hence, put simply, AI usage can only interfere with a person’s capacity for insight, never augment it.
However, that is highly problematic because it turns out that regardless of whether insight is your default problem-solving mode or a process requiring your deliberate employment, the amount of it you use predicts “how well [you] discriminate between real news stories and fake ones, as well as between meaningful statements and ‘pseudo-profound bullshit’ statements” (Kounios & Kousnios, p. 25). Insight, thus, is a cognitive superpower against all the misinformation, disinformation, and blatant falsehoods flooding our daily lives—including the classroom! —and anything that inhibits its effective usage risks our very capacity to discern what is true. Therefore, we need both insightful people as well as the more deliberate practice of insight in every element of society—again, including the classroom—if we are to find authentic solutions to the problems, both great and small, that confront us in our daily lives.
And one such problem confronting us today is a decline in people’s tolerance for healthy, beneficial risk. The research now shows that there is a direct correlation between an individual’s quantities of “aha! moments” and their degree of comfort with risk-taking, and the evidence is clear that people observed displaying greater insightfulness also display less psychological concern for potential fallibility. Indeed, the more insight these individuals employ, the more he, she, or they appear willing to engage in trial-and-error to solve problems. Thus, these individuals seem almost immune to the emotional consequences of failure, and since that is practically the definition of the growth mindset universally espoused by nearly every educator on the planet, the value of insight as a tool for learning becomes undeniable. Not that one cannot employ logic in a similar fashion. But it would seem that providing more opportunities to employ insight may help everyone approach the process of learning more effectively,[i] and that is even more reason not to employ any augmentative AI in schools that might inhibit insightful thinking. AI and “aha!” are fundamentally incompatible.
Which is why, once again, I find myself at the “end” of the perpetual news cycle still antagonistic to nearly all things AI (and associated). As an educator, I have known the value of the “aha! moments” since my very beginning in the classroom, and as an experienced educator, I know how to generate the conditions to make them happen. Indeed, as I write in the introduction to this entire project, the whole purpose behind my concept of “authentic engagement” is to enable and empower all educators everywhere to produce the insights that are the foundation of all genuine learning.
Yet, I am too much the skeptic not to wonder if the pejorative “Oh, Boomer!” may not actually apply to this aging educator and his quasi-neo-luddite sympathies. I am not going to halt the future use of AI in schools or the workplace, and I am not individually even likely to slow them down much. In fact, while I might serve as a bulwark against how my own institution employs it—maybe even influence a few others through my writing—my very finitude will eventually silence my voice regardless. What’s more, I have too much empirical data in front of me demonstrating that there is the very real possibility that none of it will matter, that there will be no one with enough IQ, CQ, and EQ to worry about my concerns in the first place.
But the latest idiotic hysteria surrounding the Hantavirus reminds me of how dangerously ignorant most of the population in this country is, and when you couple that with the current contemptuous rejection of expertise, then that danger only explodes exponentially. While I truly get why the generations I wrote about last time are questioning whether to have children—I resist the urge myself sometimes not to scream at the stories coming out of my NPR station on the radio in the morning—the reality is that some of them are having babies, and those babies are going to need a viable world in which to live. Hence, I have to remind myself that the darkness only wins if I stop shining—my own “aha! moment” I must renew each and every day.
References
Kounios, J. & Kounios, Y (March 2025) The Brain Science of Elusive “Aha! Moments.” Scientific American. Pp. 21-27.
Luse, B.; McBain, L.; & Pathak, N. (April 13, 2026) You Might Be Suffering From AI Brain Fry. It’s Been a Minute. https://www.npr.org/2026/04/13/nx-s1-5780867/you-might-be-suffering-from-ai-brain-fry.
[i] As well as potentially explain why certain individuals cling harder to a fixed mindset than others: logic may be their default mode.
Light a Candle 