AI Isn't Making Us Dumber
Every generation fears the tool that frees them
A recent MIT Media Lab study has the tech press in full panic mode. Students who used ChatGPT to write essays showed weaker brain connectivity and lower memory retention. Most couldn’t recall what they’d written. The narrative is that we’re entering a “stupidogenic society” where AI fundamentally erodes our ability to think.
Meanwhile, I’ve been using AI to correlate research across sources and iterate on ideas faster than I ever could alone. Either I’m actively destroying my brain, or the framing is missing something important.
I think it’s the framing.
We’ve Had This Conversation Before
Socrates was famously skeptical of writing. In Plato’s Phaedrus, he warns that this new technology will “produce forgetfulness in the minds of those who learn to use it, because they will not practice their memory.” Writing, he argued, would give people the appearance of wisdom without the reality of it.
The irony? We only know what Socrates thought because Plato wrote it down.
The pattern repeats. Critics worried the printing press would reduce education quality since people wouldn’t memorize passages. The telephone would destroy face-to-face communication. Calculators would eliminate mathematical ability. GPS would erode spatial reasoning. Each new tool that freed mental capacity was treated as cognitive decline.
What actually happened with calculators? Studies showed that students were freed to gain deeper understanding of mathematical concepts because they weren’t stuck on arithmetic. The cognitive effort shifted, and so did what “being good at math” meant.
What We Actually Lost
One or two generations ago, most households knew how to make clothes. My grandmother could sew her family’s clothes from scratch. I can’t (I can sew a button back on, but that’s about it).
By the “AI makes us dumber” logic, I’ve suffered cognitive decline. I’ve lost a skill my grandmother had, but I don’t feel dumber. I’ve reallocated that cognitive capacity to other things better aligned to today’s society. If I wanted to learn clothwork tomorrow, I’d have access to pattern libraries and communities my grandmother couldn’t have imagined.
This is what’s actually happening with AI. We’re not losing cognitive ability. We’re reallocating it. The real concern isn’t that AI handles certain tasks for us, but whether we’re using that freed capacity for something worthwhile or just letting it atrophy (like the Wall-E humans).
Crutch or Collaborator
The MIT study actually supports a more nuanced reading if you get past the headlines. When the “brain-only” group was later permitted to use ChatGPT, their essays showed higher creativity and stronger arguments while retaining original thinking. The sequence mattered. Think first, then augment.
This tracks with how I’ve been using AI for planning. I’m not outsourcing my thinking. I’m iterating faster on strategic ideas and testing frameworks against research rather than manually correlating data. The cognitive work didn’t disappear. It shifted to where humans are actually better.
The studies showing “cognitive decline” measure whether we remember the answer. They should measure whether we’re asking better questions.
The Real Question
Socrates was right about one thing. Writing did change how we use memory. So did printing, calculators, GPS, and now AI. Each time, we redefined what “smart” means. Each time, the panic missed the point.
I don’t feel bad that I’m no longer manually correlating data sets or aggregating interview transcripts or other manual slogs. I feel empowered that I can spend more time planning the next strategic move with those insights. The tool changed, but the thinking didn’t go away. It went somewhere more useful.
What cognitive work are you doing today that a better tool could free you from?
