My nephew understands muscles. He works out. He knows which ones need attention and which ones don’t. Recently he discovered a muscle he didn’t know he had. Reading.
Now, before anyone writes in, I appreciate that he can already read. He reads messages, websites, subtitles, social media posts, product reviews, Snapchat, Instagram, Wikipedia entries and whatever else his generation consumes during all those lingering glances at their phones. I must say I am also reaching a point where the only things I can read without strain have to be backlit.
But my nephew had just encountered a version of this muscle that felt ancient, runic, cuneiform. Yes we are talking long-form reading. Without the aid of screen or a scrollbar. The ability to sit with a hard copy of something that doesn't fit on one side of A4. And, horrors, this muscle looked FLABBY!!!
What seems to have happened is that my nephew actually found himself enjoying the Audible experience of Notes from Underground by Dostoevsky. And a dangerous thought entered his head. “I could read this in hard copy.” This was immediately followed by an even more dangerous thought. “Could I read this in hard copy?” At which point his gym logic appears to have taken over. I know this because a few days later, while helping me with some important moving business, I found him inexplicably drawn towards the Russian author section of my mother’s library, which I happen to be looking after. He had found the exact book he'd just listened to - masquerading as a shelf ornament. He was holding it and staring at it with the kind of curiosity and concern that he usually reserves for unfamiliar gym equipment. It was novel in every way.
He told me that he had also recently completed one of those online quizzes that tells you what kind of thinker you are. The result was Thomas Aquinas. My nephew was outraged. He had never heard of Thomas Aquinas. This did not prevent him taking the result personally. In fact, it annoyed him greatly to be likened to anybody, and he has since determined not to read a word of him. Poor Thomas Aquinas. Seven hundred years after his death, he has become the literary equivalent of a footballer who has picked up a red card for a tackle he never knew he’d make. He’s off the pitch.
Without warning, my nephew’s attention then turned to quizzing me about Tolstoy. “How accessible is he?” He demanded. My mouth, moving before my brain, came out quickly with, 'I think he's quite…. conceptual,' but backstage in The Suzy Logic Center, there was a slightly wide-eyed memory manager flicking through historical encounters with Tolstoy wondering how far my understanding of his writing, was going to be probed. To my relief my nephew accepted this answer without further enquiry. The question hadn't been literary, he was constructing a training regime targeting this alarmingly untoned area - and he wanted a sequence of increasingly challenging reps. Dostoevsky was his starting weight because he had liked him on Audible. Tolstoy was a name he'd heard bandied about somewhere. So he wanted to gauge strain and effort level. And would it be worth it. Aquinas obviously remained unavailable for selection. Partly because there was nothing of his available in that particular bookcase, but mostly because of The Red Card.
In the end he left with four books. Candide by Voltaire. The Uses of Diversity by G.K. Chesterton. The Dostoevsky. And a hardback copy of one of the two books his grandfather wrote on industrial relations. It struck me as an unusual reading list. Mostly because it had been assembled by a twenty-something without consulting a single algorithm.
After he left, I found myself wondering about two things. The first was whether some obscure Gen Z law had just been broken. The second was the reading question. More specifically, the reading muscle. Because my nephew wasn’t really looking for information. If information had been the goal, he already had everything he needed. Audible had provided the story. The internet could provide the summary. AI could provide the themes, the characters, the historical context and a comparison with three other Russian novelists. All before he'd squirted gym spray into his trainers in the morning. And certainly, well before Aquinas could attempt an appeal.
But that wasn’t what interested him. He didn’t want the information. He didn’t want the summary. He didn’t want the key points. He wanted the thing itself.
And that struck me as unusual. AI is becoming remarkably good at cognitive heavy lifting. It can read. Summarise. Compare. Explain. Synthesise. Draft. Interpret. Increasingly, it can do many of the things that once required sustained mental effort from us. And that’s wonderful. Right up until the point where we outsource the wrong thing. Our workout. Because the problem with outsourcing is that it often works brilliantly in the short term. The reckoning arrives later. Nobody notices they’ve lost stamina on the first day they stop going to the gym. The problem appears six months later when they realise they are now making noises when standing up out of a chair. Or for those who are younger, can no longer stand up directly from a cross-legged position on the floor…
Cognitively, when we hand over mental arithmetic to calculators and map reading to satnavs, we start to lose the stamina for those kinds of tasks. So I began to wonder what might happen if we handed more of our thinking to AI.
At which point AI calmly cracked its knuckles, limbered up and turned to its manager.
"Patrick," it said, "I think we've found ourselves a contender."
What followed was a cognitive workout regime even my nephew would have been proud of.
It began life as what I confidently described as "a quick refresh".
The WorldPrism profile report text was due for refreshment. Not replacing. Refreshing. The model was already there. The expertise was already there. Years of experience using the dimensions in learning design were already there. The task was simply to modernise the tone, tighten the language and make sure every dimension was still standing firmly in its own lane. AI seemed quietly confident we'd have this wrapped up before teatime. But with nine cultural dimensions at risk of AI 'blandification', I was already chalking my fingers, adjusting my stationery-themed headdress and limbering up for a full cognitive workout.
I gave the two of us a week.
This turned out to be adorable.
Somewhere around Day Three, the whole thing stopped feeling like editing and started feeling suspiciously like The WorldPrism Federation Wrestling Championship.
The belt was suspended over the ring. The crowd hushed. The radio commentators were ringside.
🥁 Ladies and gentlemen… Welcome to tonight's main event! Fighting out of the red corner…
Wearing sequins.
A stationery-themed headdress.
And a level of self-confidence usually associated with professional wrestlers and management consultants...
The defender of independent judgement...
The Human Red Pen!
(That's me.)
And fighting out of the blue corner…
The undefeated...
The undisputed...
The AI Heavyweight Champion!
It has read everything. Understood everything. Synthesised everything. Compared everything with everything else. Its signature move?
The Absolute Comprehensiveness Half Nelson.
A devastating hold from which few ideas escape 'unpolished'.
Before the opening bell, a commotion at ringside.
Thomas Aquinas has somehow acquired a floor pass. And approached AI. “I have some thoughts,” he said.
AI smiled politely. “I’ve got you synthesised.” This seemed to please Aquinas for approximately two seconds, until he remembered he was still banned from active play. The referee produced another red card.
DING.
ROUND ONE. TASK/RELATIONSHIP.
AI looking very confident, opens with the Absolute Comprehensiveness Half Nelson...
'...and The Human Red Pen slips free! Beautiful footwork.'
'Oh, that's AI's third attempt to polish paragraph 3'
'Bold move.'
'AI appears to think they're close enough'.
'Dangerous territory'.
The contender circles. AI advances. The Human Red Pen suddenly counters with a perfectly friendly task-focused colleague.
'AI wasn't expecting that'.
'No, it's wobbling now'…
AI recovers quickly, warmly endorses the counter and immediately embosses the paragraph.
'That's risky…'
' ...AND SHE'S THROWN IT CLEAN OUT OF THE RING!
Crowd: RED PEN! RED PEN! RED PEN!
'That's championship writing.'
Excellent separation. Great definition. Task and Relationship return to their respective corners.
FOUR WEEKS LATER…
DING.
ROUND 97: SIMPLE/COMPLEX
The Human Red Pen has retired to her corner after a bruising opening exchange. The AI Heavyweight Champion appears remarkably fresh.
He's currently sitting on a stool being fanned by a large language model while no fewer than fourteen RAGs mop his brow.
Red Pen is back on her feet. AI immediately reaches for another helpful refinement.
The crowd groans.
'Great definition in that hold'.
'Excellent separation'.
Back in the ring, AI has attempted to improve something that was already finished.
AND IT'S OUT AGAIN.
Round followed round. Far more of them than I had budgeted for. Until, finally, we had something I was properly happy with. The Nine Dimensions of the WorldPrism had survived the bout. Now they just had to survive wider expert scrutiny. The WorldPrism belt was raised. I began the traditional victory march around the ring, pumping the air in a manner entirely appropriate to a woman wearing a stationery headdress. AI applauded with impeccable sportsmanship.
And the resulting text?
Entirely ordinary.
As it should be.
The question isn’t whether AI should be in the ring. Of course it should. I like AI. I use AI. I have spent large parts of the last year willingly stepping into the ring with it. The real question is this:
Which workout are we asking AI to give us?
The Couch Potato Workout
AI, read that for me.
AI, summarise that for me.
AI, compare those for me.
AI, tell me what to think.
AI, save me the effort.
The Wrestler Workout:
AI, challenge me.
AI, give me another angle.
AI, hold the thread while I disappear down this rabbit hole.
AI, remind me where we started.
AI, tell me where my thinking falls apart.
AI, make me defend my thinking.
Watching my nephew build his reading muscles, and then spending a month in the ring with AI, left me wondering whether the delayed reckoning isn't AI itself. Because AI wasn't taking the effort from me. It was honing my reps.
Perhaps the danger lies in an unreflective surrender of the workouts we ought not to stop doing.
Some kinds of effort aren't inefficiencies to be engineered away. They're how we keep our judgement sharp, our thinking alive and our humanity in the ring.