Choking down the AI red pill

Are we now willing to accept AI is profoundly impacting the labour market?

You take the blue pill – the story ends, you wake up in your bed and believe whatever you want to believe. You take the red pill – you stay in Wonderland, and I show you how deep the rabbit hole goes.
Morpheus, The Matrix, 1999

More employers plan to replace workers with AI within the next year, according to a September [2025] report by Resume.org. Nearly 3 in 10 companies said they’ve already replaced jobs with AI, and by the end of 2026, 37% expect to have replaced jobs with AI. “AI adoption is going to reshape the job market more dramatically over the next 18 to 24 months than we’ve seen in decades,” Kara Dennison, head of career advising at Resume.org, said.
Carolyn Crist, HR Dive, 22/9/2025

Job openings have mostly tracked the rise and fall of the stock market. Nothing like the current divergence between equity returns and job openings has ever occurred in the history of the Job Openings and Labor Turnover Survey.
Derek Thompson, 23/10/25

[Job] advertisements were 4.3 per cent lower than a year ago. This confirms the anecdotal message that there is a significant change in labour market conditions, which is yet to be seen in official statistics… The slowing job market will be accelerated in the next 18 months as artificial intelligence and enhanced computing slash the number of people working in a wide area of professional activities… There is great frustration building among what is a generation of people who suddenly find they are in a world they did not expect.
Robert Gottliebsen, The Australian, 4/11/25

For the past year, the leading narrative around AI has been: Don’t worry, ChatGPT won’t take your job… Many CEOs have tried to reassure workers that AI will enhance, not replace, their jobs… Yet, the headlines tell a different story… So far this year, U.S. employers have announced 946,000 job cuts, the most since 2020… The question now isn’t whether AI will take jobs, it’s how you’ll protect yours.
Scott Galloway, Prof G Markets, 3/11/25

From the perspective of the common person, the elites are the equivalent of artificial intelligence — a collection of people who, in aggregate, reason using methods and techniques which are foreign to them. Moreover, recent years and decades have shown that there is indeed a misalignment. Between deindustrialization, automation, mass migration, globalization, and financialization, it’s clear that the cognitive elites will bend society in a direction that will disproportionately favor their own class, even in the scenarios where it disproportionately negatively impacts everybody else.
Rohan Ghostwind, Brackish Waters, Barren Soil, 3/11/25

Before getting down to business, let me explain what I bring to the table when opining on AI-driven labour market disruption.

My competitive advantage over the many AI doomers who are vastly more intelligent than I is that I’ve spent a couple of decades working in or around a collapsing industry. In contrast, many who share my fears – Sam Altman, Dario Amodei, Mo Gawdat, Demis Hassabis, Geoffrey Hinton – have, or had, secure, well-remunerated, high-status roles at hugely profitable tech companies.

AI had nothing to do with the collapse of the legacy media. That was caused by Google, Facebook and YouTube gobbling up the abundant advertising revenue newspapers, radio stations and TV networks once had a monopoly on.

But technological obsolescence is technological obsolescence, regardless of whether the technology is a printing press, power loom, combine harvester or social media network. And human reactions to being rendered obsolete don’t appear to change much.

Since the publication of Elisabeth Kübler-Ross’s On Death and Dying in 1969, we’ve had a handy framework for understanding how people typically react to profound change. Be it death itself, or somewhat less final ‘little deaths’, such as the passing away of a life stage, relationship or career.

Given my occupational background, I’m familiar with the five stages of grief. I’ve observed many one-time colleagues and industry peers cycle through denial, anger, bargaining, depression and acceptance. I’ve transitioned through these powerful and judgment-clouding emotions myself.

So, when ChatGPT dropped, almost three years ago to the day, I made two predictions to myself that have subsequently turned out to be accurate.

As a business and tech content creator, it was obvious to me that employers would aggressively slash labour costs using AI as soon as the technology matured enough for this to be feasible.

I also believed it would be a long time before a critical mass of people accepted the evidence of their lying eyes about AI-driven labour market disruption.

You just keep telling yourself that
I may no longer be a print journalist, but I remain, at least for now, a professional storyteller. One with a keen interest in the narratives that individuals, classes and societies embrace to make sense of the world.

With reference to Kübler-Ross’s schema, let’s list the narratives that have been cycled through since ChatGPT’s launch on November 30, 2022.

Before proceeding, it should be noted that individuals, groups and societies rarely move through the five stages in an orderly and sequential manner, and that it’s common to experience two or more stages simultaneously. For instance, an individual will often both insist something isn’t happening and that if it is – which, of course, it definitely isn’t – they can devise and execute a cunning plan to stop it happening.

The gleeful manner that many media outlets and their audiences seized on a recent MIT report that claimed, as the headlines screamed, “95 per cent of AI pilots fail” demonstrates just how alluring denial remains, even at this late hour.

I imagine many AI sceptics didn’t scroll much past the headlines and, even if they did, didn’t find an alternative perspective. Even if an alternative perspective was provided, they no doubt dismissed it and once again reassured themselves AI would ultimately prove to be an overhyped nothingburger.

I’m unlikely to convince anyone still stuck in the denial stage the coverage of the MIT report was lacking, but here goes.

The MIT report never claimed AI was overhyped or that it wouldn’t ultimately vaporise many jobs.

It did find 95 per cent of generative AI pilot programs failed to deliver the rapid revenue growth and/or cost reductions that were hoped for. But if you got in a time machine and went back to the late 1990s, you’d probably find 90 per cent of internet pilot programs also failed to deliver on expectations. It would have been unwise not to invest in Facebook (est. 2004) or YouTube (est. 2005) on that basis.

The main issue with the technology, according to the report, was that “most AI tools don’t learn and don’t integrate well into workflows”.

That shortcoming is precisely what AI agents promise to address. As the name suggests, these agents can learn from new data and interactions with humans, be deeply integrated in business processes, and act independently.

Joining the reality-based community
In the textbooks, the stages are: Denial, Anger, Bargaining, Depression, Acceptance. But as white-collar workers witness AI collapsing the labour market, I suspect the sequence will be: Denial, Bargaining, Depression, Acceptance, Anger.

Stage One: Denial
It’s hard not to be impressed by AI, especially the first few times you encounter it. The first iteration of ChatGPT seems primitive in retrospect. But many, including me, were gobsmacked that you could give an AI platform instructions in vernacular English, aka natural language, and have it summarise documents, draft work emails, explain complex topics, or compose poems.

But to the present day, many smart people insist there’s nothing particularly impressive about Large Language Models, and they’re just a “glorified autocomplete” or a “stochastic parrot”.

Stage Two: Bargaining
Those in the bargaining stage like to nod sagely and proclaim, “AI isn’t going to take your job, but someone using AI might!”

Early adopters who embrace AI tools will be more attractive to employers than late-adopting technophobes. But there are two issues with the being-someone-who-uses-AI strategy.

First, a substantial investment of time and effort is required to develop the advanced AI skills employers want.

Second, even if you’re inclined to, for instance, spend hundreds of tedious hours mastering programming languages such as Python, SQL, HTML and CSS, you’re still likely to be replaced by an AI agent in the nearish future.

In short, while someone using AI might take your job initially, AI will take it from them shortly afterwards.

Stages Three and Four: Depression and Acceptance
These things are impossible to quantify, but it appears a critical mass of white-collar workers have passed through the denial and bargaining stages and are currently in a discomfiting limbo between depression and acceptance.

Despite much-publicised layoffs throughout 2025 at the likes of Accenture, Alphabet, Amazon, Meta, PwC and Salesforce, most knowledge workers haven’t yet been abruptly ushered from their cubicle by a security guard. Thus far, most businesses have only instituted hiring freezes, rather than reaching for the scythe.

However, lots of knowledge workers now know someone who has lost their job or business because of AI. Plus, for anyone willing to notice them, there are plenty of other red flags being unfurled. Such as rising unemployment rates (especially for graduates), employers leveraging their strengthening bargaining position (see return-to-office mandates and increasingly arduous hiring processes), fewer job opportunities, and more ‘job hugging’.

For anybody with even a rudimentary understanding of how businesses operate, there’s also the small matter of the world’s most successful, prescient and ruthless CEOs and entrepreneurs betting the farm on AI. (Gartner estimates worldwide spending on AI in 2025 will total around US$1.5 trillion this year and will reach US$2 trillion next year.)

Amazon, Alphabet, Microsoft and Meta have ploughed enormous sums into AI. Sooner rather than later, they’ll need to start making a return on their investments. When the tech bros and C-suiters talk vaguely but upliftingly about “finding greater efficiencies” or “freeing staff up from routine tasks”, what they are referring to is replacing an employee pulling down $100,000 with a $1,000 per annum AI subscription.

Chart from Derek Thomson's post, ‘Is This the New ‘Scariest Chart in the World’?

Stage Five: Anger
Even in the best of circumstances, societies would struggle to cope with the profound disruption to labour markets that’s now starting to unfold.

These are far from the best of circumstances for Anglosphere societies. In recent decades, these societies have been atomised and fractured along class, gender, generational, geographical and ethnic lines.

It’s challenging to forecast what happens when you pour the fuel of rising unemployment on the dumpster fire of fragmented societies. But it’s inevitable that a tidal wave of layoffs will force those hurled on the scrapheap to abandon the status- quo-preserving stories they’ve long told themselves and embrace radically different narratives.

What comes after Stage Five?
I’m writing this a few hours after New Yorkers elected a youthful Muslim socialist as their mayor, a week after Tucker Carlson platformed Nick Fuentes, a month after Charlie Kirk was assassinated, a few months after both the UK and Australia hosted well-attended anti-mass immigration marches, and nine months after Donald ‘Literally Hitler’ Trump began his second term.

The pace of technological and political change will only accelerate in 2026.

I can only guess as to exactly how that will play out. But I’m certain those who’ve swallowed the red pill will be better prepared for whatever is coming than those who insist on downing blue pills and continuing to believe what they want to believe.