Unintended AI Consequences #3: Gutting the Cognitive Class
In the 80s and 90s, there was a pretty clear upgrade path if you were “one of the bright ones”.
You did well at school, maybe went to uni, learned to string a sentence together or solve a spreadsheet, and you were basically set. Not rich, but upwardly mobile. A professional. The sort of person who could get a mortgage, buy a Golf GTI, and tell their kids, “Work hard and you’ll do better than us.”
The assumption was simple: If you were smart and you could work with ideas instead of a shovel, the economy would look after you.
That was the deal. Especially in places like London. Every new wave of “bright young things” – grads, junior consultants, copywriters, analysts, planners – pitched up in the city, sold their brains, and climbed.
Fast-forward to now, and that deal is quietly collapsing.
The same “cognitive class” that’s done well out of the last 50 years – traditional middle-class knowledge workers – are suddenly standing on thinner ice. Wages are flat, careers are choppy, everything looks like a gig, and a lot of that precious cognitive work is being gobbled up by machines that just don’t get tired.
And that’s one of AI’s nastier, unintended consequences:
It’s not just automating low-level jobs. It’s hollowing out the rungs of the cognitive ladder that used to make the middle class feel safe.
Who are we talking about?
When I say “cognitive class”, I mean:
• The grads who went into media, marketing, consulting, law, finance, comms, design, policy.
• The office-based professions that live on PowerPoints, Google Docs, Slack channels, and Zoom.
• The “bright young things” who were always told they were destined for “good jobs”.
In other words: the people whose main asset is their ability to:
• process information,
• write and speak clearly,
• analyse,
• plan,
• persuade.
For the last few decades, this has been a great place to be. The world shifted from heavy industry to services, from factories to offices, from brawn to brains. If you could think, present, and bullshit in a convincing way, you were golden.
There was even a kind of moral story attached to it: “We deserve this. We studied. We worked late. We learned Excel. We’re the ones making the hard decisions.”
Then along comes AI and, with a straight face, basically says:
“That thing you get paid £250 a day for? I can do the first 90% in about four seconds.”
Awkward.
The invisible bargain that’s breaking
For my generation – Thatcher kids who got out – the cognitive bargain went like this:
1. You invest heavily in education. Time, money, emotionally. Maybe you take on big debt.
2. You get access to higher-paying, more secure jobs. They might be dull, but they’re safe-ish. Career ladders, pensions, stability.
3. You convert that into a middle-class life. House, kids, holidays somewhere sunny, maybe a German car on the drive.
Except, now AI comes in at Stage 2 (or maybe earlier) and quietly wrecks the economics.
Because if you’re a company looking at your P&L, the logic is brutal:
• Why hire 5 junior analysts when 1 mid-level with AI tools can do similar output?
• Why pay copywriters, designers, planners full-time when you can externalise risk to a gig pool and plug them into AI?
• Why invest in grad programmes when you can wait and buy mid-levels later – after the market has chewed them up a bit?
The result isn’t some sudden robot takeover. It’s something slower, nastier, and harder to protest:
The early rungs on the professional ladder get thinner, more precarious, and worse paid – just as the price of entry (degrees, debt, living costs) has gone up.
Gutting from the middle, not just the bottom
We had this comforting idea that technology automates the boring low-end work and humans just “move up the value chain”.
The problem with AI is that a lot of “value chain” work is… typing, researching, and shuffling information around. Exactly the stuff junior and mid-level knowledge workers do all day. And AI is also racing up the value chain.
Watch what’s happening in real time:
• In marketing and media
The £35k copywriter, social manager, or junior planner now competes with tools that can spit out 20 variants in minutes - all sense-checked and with the correct tone of voice. You still need humans – but fewer, and more senior.
• In consulting and strategy
Decks that used to take a week of analyst graft can be templated and partially generated. Fewer spreadsheet monkeys, more “partner” types who sell and frame the work.
• In design and content production
AI tools are churning out first passes for logos, layouts, video edits, transcriptions. The grunt work that trained juniors is being eaten from underneath.
What does that do?
• It compresses the middle – junior and mid roles – and inflates the extremes: a handful of high-paid stars at the top, and a sprawling mass of freelancers, contractors, and gig workers at the bottom.
• It turns “learning years” into “fighting for scraps” years. The grind that taught you the craft becomes a scramble to keep your day rate above your rent.
Flat wages, fake freedom, and the gig illusion
Outwardly, this is sold as liberation:
• “You’re a consultant now, not an employee.”
• “Portfolio career.”
• “Be your own boss!”
In reality:
• You’re juggling three clients, two overdrafts, and one landlord.
• You have no benefits, no sick pay, no pension, and an algorithm somewhere deciding whether you get the next job.
• You’re in “the creative industries” but you can’t afford a flat anywhere near the city that houses them.
AI amplifies this because it:
• Commoditises a lot of cognitive work (there’s always someone who’ll do it cheaper, faster, or semi-automated).
• Obscures the human contribution (clients see the output, not the juggling behind it).
• Erodes pricing power (if your client thinks “a bot could do this”, your negotiating position is already weaker).
We’ve turned a whole swathe of previously decent jobs into a kind of white-collar zero-hours economy. Same laptop. Same coffee shop. Same “I work in digital”. Very different level of security.
Debt on the way in, insecurity on the way out
Here’s the cruel bit.
We’ve told an entire generation:
“Borrow tens of thousands for a degree, move to an expensive city, grind hard in your 20s – and it’ll pay off long-term.”
That made sense when:
• There were stable jobs at the end of it.
• Wage growth roughly tracked experience and inflation.
• You were actually accumulating some kind of social and financial capital.
Now:
• Degrees are more expensive.
• Rents in London (and other hubs) are insane.
• And the jobs at the end are less secure and more easily automated.
So you get:
• A big cognitive class with fragile finances.
• People delaying kids, homes, and long-term commitments because their work lives feel like a treadmill on ice.
• A simmering sense of “I did everything right – why do I still feel shafted?”
If you want a breeding ground for resentment and political volatility, that’s it. Not the very poorest – they’ve always been shafted. The people just above them, who were told they were joining the winners’ club, only to find the bar’s been moved.
The emotional gutting: status anxiety at scale
The economic hit is one thing. The psychological hit is worse.
For the cognitive class, work isn’t just money. It’s identity.
• “I’m a strategist.”
• “I’m a lawyer.”
• “I’m a journalist.”
• “I’m a designer.”
If a machine can do a large chunk of what you do – or at least what you thought made you special – it’s not just your income on the line. It’s your story about yourself.
You see this already:
• People loudly proclaiming AI is “just a toy” or “only good for basic stuff” while clearly being rattled.
• Professionals more invested in defending the purity of their craft than in evolving it.
• A rising hum of anxiety: “What if my job is just pattern-matching with a nicer haircut?”
That’s what “gutting the cognitive class” really looks like:
• Less like robots marching in, more like a slow, steady hollowing out of confidence, status, and security.
Is there any way this doesn’t end in a clusterfuck?
I’m not a doomer by default – I’m temperamentally optimistic. But optimism isn’t pretending the hit isn’t real. It’s asking: what can we do with this mess?
A few hard truths and possibilities:
1. The old deal is not coming back
The “study hard, get a degree, walk into a stable middle-class job” bargain is gone. That doesn’t mean all is lost; it means we have to stop lying about it.
Parents, schools, universities, governments – they’re still selling a story that belonged to the 1990s.
AI is just accelerating us all away.
2. The value of how you think goes up as the value of what you produce goes down As AI eats more of the production work, the premium shifts to:
• Framing problems.
• Asking better questions.
• Judgement, taste, ethics.
• Building trust and relationships.
• Being the person people want to work with, not just the person who can operate the tools.
The cognitive class will survive – but not by clinging to the idea that their worth lies in being better at manual analysis, drafting, or slide-making than a machine.
3. Ownership and leverage matter more than degrees
If you’re selling hours in a marketplace where AI can undercut you, you’ll get squeezed, nay, screwed.
If you:
• own the relationship with the client,
• own the product,
• own the distribution,
• or own the IP
…you’re in a different game.
One of AI’s unintended positive consequences is that more people could build and own things: products, mini-agencies, small but global businesses – if they stop thinking like employees and start thinking like small-scale capitalists.
Easier said than done, but still true.
4. We will need policy, not just hustle culture
I’m all for personal responsibility and seizing opportunity, but let’s not pretend the market will gracefully handle this on its own.
If we don’t want an angry, disillusioned cognitive class drifting into nihilism, we’re going to need:
• Smarter safety nets.
• Rethinking education (shorter, cheaper, more modular, more honest).
• Tax and regulatory frameworks that don’t just shovel more chips to the already loaded.
So where does that leave the “bright young things”?
If I were 19 today, turning up in London with a laptop and some brains, I’d be looking at AI and thinking two things at once:
1. “Shit, this could absolutely undercut me.”
2. “Also, this might be the only way I get out ahead.”
The cognitive class is being squeezed, yes. But that same squeeze is creating weird, messy, new spaces where the brave and slightly unhinged can build interesting lives:
• Tiny teams with massive leverage.
• Niche experts amplified by AI, not replaced by it.
• People using AI as an exoskeleton for their judgement, not as a crutch for lazy thinking.
The risk is real. The gutting is already underway.
But the story’s not finished. AI is hollowing out the old, comfy version of the cognitive class – the one that believed a degree and a job title was a lifetime safety blanket.
The unintended consequence is a shake-up that will hurt a lot of people and throw up new shapes of success that look nothing like the 9–5 office ladder. Thankfully.
The real divide in the next decade won’t be “educated vs uneducated”. It’ll be:
Those who adapt to being a human with AI vs those who cling to a world where their brain was the only machine in the room.


















