We are living through a strange experiment in human cognition. Artificial intelligence has moved into our lives with astonishing speed. It finishes sentences, synthesises information, generates ideas, writes first drafts and increasingly takes on work that once required sustained human attention.
Much of this feels like liberation. Things that used to take hours can now take minutes. Blank pages disappear. Information arrives before we have finished articulating the question.
And yet, underneath the exhilaration, I keep noticing another feeling. A kind of cognitive unease. Am I still thinking as deeply? Am I becoming less patient with difficulty? Is my own voice getting harder to hear inside all this generated fluency?
Perhaps most unsettlingly, what happens to human capability when the machine becomes very good at doing the things through which we once developed it?
That question has started appearing in the research too. One widely discussed MIT study followed people completing writing tasks with and without large language models. The findings were preliminary, and the study should not be treated as a final verdict on what AI is doing to our brains. But it raised an important possibility: when people outsourced more of the cognitive work, researchers observed differences in engagement, recall and ownership of what they had written.
The internet understandably seized on the most frightening interpretation. AI is making us stupid.
Perhaps. But I wonder whether that conclusion is too small.
Throughout human history, new technologies have repeatedly changed which capabilities we need to exercise directly. Writing reduced the need to hold entire cultures in memory. Calculators changed how we perform arithmetic. GPS changed how we navigate. Search engines changed what we need to remember in order to retrieve information.
Every extension of human capability has also involved some degree of relinquishment. Something moves outside us. And when it does, the more interesting question becomes: what grows in the space that remains?
This is where the conversation about AI begins to feel much bigger than productivity. Perhaps we are not simply witnessing the automation of tasks. Perhaps we are witnessing a redistribution of intelligence.
Machines are becoming extraordinarily capable at processing, generating, retrieving and recombining information. That does not mean human cognition becomes irrelevant. It may mean the centre of gravity moves.
Away from simply producing answers and towards knowing which questions matter. Away from information and towards judgement. Away from fluency and towards discernment. Away from knowing more and towards understanding what deserves our attention.
And perhaps, somewhere underneath all of that, towards capacities that are harder to separate from being a living human being: embodiment, relationship, meaning, intuition, responsibility and moral imagination.
Not because these capacities are mystical or uniquely pure. Humans are biased, irrational and frequently terrible at all of them. But because intelligence in the real world has never existed independently of a body, a history, a community or a set of consequences.
We do not make important decisions from information alone. We make them while tired, while afraid, while in love, while protecting something, while wanting to belong, while remembering what happened last time, and while trying to imagine what our choices might mean for another human being.
Our intelligence has always been entangled with life. AI makes that easier to see.
For centuries, particularly in the West, we have tended to privilege cognition as though thinking were the highest expression of being human. I think, therefore I am.
But the arrival of machines capable of extraordinary feats of cognition unsettles that hierarchy. If a machine can write the report, solve the equation, generate the strategy and synthesise the literature, perhaps the question of human value cannot rest entirely on our ability to produce cognitive output faster than something else.
Perhaps our definition of intelligence has been too narrow all along.
I feel, therefore I am. I relate, therefore I am. I choose, therefore I am. I care, therefore I am. I make meaning, therefore I am. I live inside the consequences of my decisions, therefore I am.
None of this means we should cheerfully surrender our ability to think. Quite the opposite. There are forms of friction worth protecting.
The effort required to wrestle with an argument. The discomfort of not knowing. The patience involved in forming an original idea. The cognitive work of deciding what we actually believe before asking a machine to articulate it for us.
If AI removes every difficulty, it may also remove some of the conditions through which capability develops. And that distinction matters.
Because there is a difference between friction that wastes us and friction that forms us.
The tedious administrative task that consumes half a day may be worth automating. The difficult conversation probably is not. Searching through fifty documents for one piece of information may add little to our humanity. Sitting with competing values before making a decision might. Writing the first draft of a routine email may not matter very much. Finding your own language for something you deeply believe probably does.
Perhaps that is the deeper opportunity in this moment. Not to resist every form of cognitive outsourcing, nor to hand ourselves over entirely to the machine, but to become more discerning about which parts of being human we are willing to outsource.
Because atrophy is not inevitable. But neither is evolution.
Evolution requires pressure, adaptation and response. The technology provides the pressure. What happens next depends, at least partly, on us.
We can use these systems to make ourselves more passive, more derivative and less willing to struggle with complexity. Or we can use them to remove the work that diminishes us and make more room for the work that enlarges us.
More thinking at the level of systems. More courage at the level of decisions. More attention at the level of relationships. More imagination. More responsibility. More presence.
Perhaps the future of human intelligence will not be defined by our ability to compete with machines at the things machines do brilliantly. Perhaps it will be defined by whether we become more capable of doing the things intelligence was always for.
Knowing what matters. Choosing what we value. Understanding one another. Seeing consequences beyond ourselves. Creating meaning from an overwhelming world. And knowing when not to take the easiest answer simply because it is available.
So yes, I take the fear of atrophy seriously. I feel it in myself sometimes.
That small temptation to hand over the difficult bit. To ask before I have wondered. To generate before I have imagined. To reach for the answer before I have spent enough time inside the question.
But perhaps that feeling is useful. A signal. A reminder that not every capability worth preserving will preserve itself.
The future is not asking us to reject artificial intelligence. It may be asking something far more demanding.
To become more intentional about the intelligence we choose to keep alive in ourselves.
With love and curiosity,
Becs x


