Human evolution is usually pictured as a march forward.
We learned to walk upright, make tools, control fire, develop language, build civilizations, and eventually create machines capable of extending our physical and intellectual abilities. Now we have created artificial intelligence capable of writing, analyzing, planning, programming, explaining, recommending, and increasingly helping us make decisions.
Look at the image above – after the modern human comes AI-assisted human.
But what comes after that?
That question interests me because humans have always used technology to reduce the amount of work we have to do. The wheel reduced physical labor. Writing allowed knowledge to survive outside human memory. Calculators took over arithmetic. GPS eliminated much of the need to remember routes. Search engines gave us instant access to information we needed, for research, or to remember.
AI takes this progression somewhere different.
It doesn’t merely store information for us. Increasingly, it can think through problems for us. Could that possibly be a step backwards instead of forward?
We already ask AI to summarize documents, write emails, research purchases, create schedules, analyze data, develop arguments, program computers, explain complicated subjects, brainstorm ideas, evaluate alternatives, and recommend courses of action.
None of that is inherently bad. Used well, AI may become one of the greatest intellectual tools humanity has ever created.
But what happens when assistance becomes dependence?
Many feel we have already started outsourcing our minds
This is not new, researchers were asking related questions long before ChatGPT and other now-popular LLMs and AI models.
In the 2011 Science paper “Google Effects on Memory: Cognitive Consequences of Having Information at Our Fingertips” Betsy Sparrow, Jenny Liu, and Daniel Wegner conducted four experiments examining what happens when people expect information to remain available electronically. Participants were less likely to remember the information itself and better at remembering where it could be found. The researchers described the Internet as a form of external or “transactive” memory.
It rings true to me. Think about your own life. How many telephone numbers do you know compared with thirty years ago? When was the last time you memorized directions before driving somewhere unfamiliar? When was the last time you went to the library to research something?
We didn’t necessarily become less intelligent. We simply changed what we expected our brains to remember.
And that cognitive offloading can actually be useful. Benjamin Storm and Sean Stone’s 2015 study, “Saving-Enhanced Memory: The Benefits of Saving on the Learning and Remembering of New Information” found that saving information externally could improve people’s ability to remember new information afterward. In other words, freeing the brain from one task can make cognitive resources available for another.
The question isn’t whether outsourcing cognition is automatically harmful.
It’s what we do with the capacity we free up. Maybe more importantly, do we do anything with it?
Suppose I use AI to organize research so I can spend more time evaluating it, challenging assumptions, and developing my own ideas. AI has reduced my workload while potentially allowing me to think more.
But suppose I ask AI what the evidence means, what conclusion I should reach, what I should believe, and then accept its answer.
Something very different has happened.
An intriguing 2025 MIT Media Lab preprint went further. “Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task” compared participants writing essays with an LLM, a search engine, or no external tools while researchers measured brain activity using EEG. The brain-only group showed the strongest and most distributed connectivity, search users were intermediate, and LLM users showed the weakest connectivity during the task. The authors also reported differences in recall and feelings of ownership over the essays.
Though that study was relatively small, it reinforces what many intuitively already suspect: The brain, like muscles, becomes stronger with usage.
The brain we exercise is the brain we build
Our brains remain capable of changing in response to experience. Learning and repeated practice strengthen skills and alter neural networks. We don’t exercise our brains by simply possessing them any more than we exercise muscles by possessing those.
Consider what would happen if children grow up rarely needing to struggle with a question.
An AI tutor knows the answer.
Writing is difficult? AI drafts it.
Can’t understand the argument? AI summarizes it.
Can’t decide which position makes sense? AI evaluates it.
Unsure what career to choose? AI knows your abilities and the employment market.
Now consider 5, 10, or 100 generations relying on AI for those types of questions.
Eventually, the “system” may know your finances, health, relationships, education, personality, preferences, history, strengths, weaknesses, and previous decisions.
Should I take this job?
Ask AI.
Should I move?
Ask AI.
Who should I date?
Ask AI.
What should I believe about this issue?
Ask AI.
And if its recommendations repeatedly produce better results than our own decisions, refusing to consult it may eventually seem foolish.
Now, look again at the evolution image
The important question may not be whether humanity biologically evolves into something different anytime soon. Evolution doesn’t work simply because we stop practicing a skill for a few years, decades, or even generations. So biologically, we are not going to change quickly… But,
Something potentially more immediate could happen.
We could culturally devolve.
A person doesn’t need to lose the biological capacity for navigation to become unable to navigate without GPS. We don’t have to lose the capacity to remember telephone numbers if we simply never bother learning them.
Could reasoning eventually work the same way?
Imagine children raised by parents who rarely make consequential decisions without AI, taught by AI tutors, working alongside AI systems, and eventually raising children of their own under even more capable systems.
Perhaps humanity remains biologically capable of independent thought while gradually losing the habit, confidence, education, and desire required to practice it.
Or perhaps exactly the opposite happens.
Maybe AI removes routine intellectual labor and allows us to devote more of ourselves to creativity, philosophy, discovery, relationships, science, and problems previously beyond individual human capacity.
The technology itself doesn’t necessarily decide between those futures.
How we use it might.
That’s why I keep returning to that final question mark.
AI may give us answers faster than any human being could ever find them. It may solve problems that would otherwise consume years of our lives. And perhaps that is exactly what we should want from it.
But human life has never been entirely about reaching the destination.
Sometimes the struggle to find an answer changes us. We learn from being wrong, reconsider what we believe, discover questions we didn’t know to ask, and occasionally find something completely different from what we originally went looking for.
If AI simply gives us the destination, even if we ask for it, do we lose something by skipping the journey?
AI could very well make thinking unnecessary. With enough time and generations, maybe even less capable of thinking for ourselves. The real question is: Do we want it to, and miss that journey that is uniquely human? The answer to that is not in the technology itself, but in when and how we choose to use it.
About This Seed
This question grew out of the short story, A Friendly Caution, and made me wonder about a quieter risk of AI. I'm less concerned here about machines suddenly taking away our ability to make decisions than about us gradually handing that ability over because it's easier. We already outsource memory, navigation, calculations, Alexa turning on and off lights and appliances, Facebook's algorithms telling us what we are going to see in our stream, and countless everyday tasks. What happens when we begin outsourcing judgment too? Discretion? Morality?
I'd love to see where others take this. Challenge the premise, argue that AI will make us smarter, imagine a generation raised without independent decision-making, create artwork showing that future, or explore what happens when someone finally decides to stop asking the machine... or a world where we have forgotten the world where we had to do everything ourselves. ~ Rick