AI.
The elephant in the room.
I think most of us have had some fun with AI tools at some point. Early in 2025, I was quite fascinated with AI and enjoyed learning about it. It is, understandably, a very divisive topic among writers and illustrators, and while this has my attention, I am also deeply concerned about the psychological and developmental impact on children.
Watching AI trends over the last year, as well as following the emerging science looking at the impact it has on our lives and brains, has turned my initial fascination into one of concern. While I am not a full-blown hater of the potential of AI, I have concerns and believe we should approach it with some skepticism and eyes wide open.
Most importantly, I believe educators, parents, and those who advocate for child welfare need to be leading the charge to press our lawmakers to slow down the adoption of generative AI and AI chatbots and debate the questions of ethics and safety before it’s too late.
AI is an experiment we are participating in on a grand scale, and in every experiment, there is a disclosure you must sign and agree to in order to join. It tells you all the risks and the things that could go wrong, and you agree that they are worth it. With AI, there is no disclosure, and we are all learning as we go. We are the AI lab rats participating in a psychological experiment.
There have been some fun AI trends that many people participated in, such as the 'roast my Easter photo' or 'turn me into a Christmas elf.' These feel harmless and fun. I am not a cybersecurity professional, so I will stay out of the discourse of whether this is a good idea, but most of us have done it and had a good laugh.
However, these types of fun, harmless-feeling AI party tricks can dull our sense of wariness regarding much greater AI themes.
One area where we know AI has a significant negative impact is in using it for brain short cuts.
Anyone else in their 40's or older remember a parent who had just been put on the spot by a homework question they didn't know the answer to, saying, "Go look it up in the encyclopedia/ dictionary?" I can remember rolling my eyes harder than the Robert Downey Jr. meme whenever I would hear this. But, like in so many things, Mom and Dad were right! They also had no idea of the answer, and the internet didn't exist, but in some ways I think that helped protect our young brains.
We would trudge over to the encyclopedia and try to utilize our poor alphabetical order skills to find whatever random keyword we were looking for and then lay the behemoth book down on the coffee table and proceed with turning the super-thin pages without ripping them. After an hour of looking at every picture and interesting topic on the way to our keyword, we would realize it was dinner time and that we still hadn’t finished our homework. Good times!
How is this different from "Ask Chat" and "Google it" generation?
In the 90's, with our old school Britannicas in hand, there were many steps required between question and answer, which required a lot of brain motivation to reach. These steps produced the good kind of dopamine spike, the kind that is earned through work and the delay of gratification. AI and Google(now that it is AI-powered by Gemini) remove the resistance and hand it to us on a silver platter without any investment or mental engagement. Question to answer without the middleman of our brains.
When we do not experience delay of gratification or the dopamine reward of the hard earned accomplishment and instead outsource these brain tasks to someone or something else, we weaken the part of our brain that is capable of doing hard mental tasks and strengthen the part of the brain that demands hedonistic ease.
Outsourcing our mind tasks sets us up for weak frontal lobes and to become perpetrators of the dreaded Dunning-Kruger Effect. This is the psychological phenomenon where people of actual low intelligence think they are smarter than they are because they have access to just enough info about a topic to fool people into thinking they are knowledgable and trustworthy. When in truth, experts in their fields are typically the people saying how much they still have to learn. Those suffering from the Dunning-Kruger Effect can give dangerously bad information with a great deal of confidence. This is AI. A roulette wheel of just enough right information to seem like an authority, presented with the confidence of an expert while being peppered with dangerously bad information.
Why should we care about the frontal lobe?
The frontal lobe is the part of the brain responsible for executive functions. It helps us plan, focus, solve problems, think ahead, and (in my opinion, most importantly) control impulses.
In other words, when we have a weak frontal lobe, we have poor attention and focus, declines in critical thinking and our ability to see long-term harm or gain, reduced emotional regulation, and the impulse control of a 2-year-old. Our frontal lobes are not fully developed until around the mid-twenties, which is why so many young people make dangerous, impulsive decisions.
AI can help us be better by giving us the information and skills (carefully vetted, please, with sources cited) for learning new things. Asking AI to help you fix your broken dryer, followed by you then doing so. That is great! You developed a new skill with AI assistance, where in the past you would have called a repairman. But if we are outsourcing tasks we would otherwise be doing ourselves(cognitive offloading) with our own brain cells, then that could be a problem.
What does this have to do with children and literacy?
According to Risko and Gilbert (2016), cognitive offloading refers to using physical actions or external aids to change the information-processing demands of a task to reduce internal mental effort. Examples include writing reminders, setting alarms, or using tools like calculators and navigation aids instead of holding all details in mind.
While some cognitive offloading can improve focus (setting alarms rather than tracking the time as you work) one of the key concerns around cognitive offloading is that while it makes tasks feel easier in the moment, it also can reduce how often people fully engage their own mental abilities. The authors explain that when individuals regularly rely on external supports instead of using memory, planning, or problem-solving internally, those internal processes are used less frequently. Over time, this habitual offloading could mean fewer opportunities to practice and strengthen core cognitive skills such as working memory, attention, and self-monitoring. They caution that consistently choosing external aids when internal processing is possible may come with costs that are not immediately visible. In particular, they raise concerns that people may offload based on perceived effort rather than actual necessity, which could lead to an over-reliance on tools and a gradual weakening of confidence in one’s own cognitive abilities.
If we are allowing our children to use a technology to do tasks which we know are necessary for strengthening their developing frontal lobes decades before that part of the brain is fully formed, we are setting them up for possible arrested neurological and emotional development, and we are removing the hard earned dopamine that will help motivate them to learn the hard tasks of reading and writing.
Even more concerning is that a once rare early-onset type of dementia characterized by shrinkage and death of the frontal lobe, called Frontotemporal Dementia, is on the rise worldwide. While there isn't yet a definitive understanding of the cause, we do know that certain behaviors cause the frontal lobe to be stronger by increasing brain activity and new cell growth, and some that weaken it.
Reading and writing skills are developed through hard work and are maintained in a 'use it or lose it' sort of way through continued mental challenge. While most people don’t have a modern set of encyclopedias to send their kids to, it is just an important reminder that in the areas where we can encourage a deeper engagement with the material, the more likely we are to generate an upward spiral of healthy neurological growth.
AI may be part of our future but it doesn't mean we are helpless to it and it certainly doesn't mean that it needs to be in the hands of children whose brains are developing who have no understanding of the potential consequences that might stem from its use. Maybe a wait and see rather than a dive head long into it would be a wiser approach.
Urso, D., Giannoni-Luza, S., Brayne, C., Ray, N., & Logroscino, G. (2025). Incidence and prevalence of frontotemporal dementia: A systematic review and meta-analysis. JAMA Neurology, 82(9), 930–940. https://doi.org/10.1001/jamaneurol.2025.3307