How do the politics of AI-driven mass automation play out?

After half a century of winner-take-all neoliberalism, can anybody even imagine a kinder, gentler society?

Image courtesy of MOeinium

How did past societies end their disintegrative periods? A critical step is reversing the root cause of crisis—the perverse wealth pump that takes from the common people and gives to the elites… Good times breed complacency among the governing elites. During integrative phases they become tempted to rejiggle the political economy in ways that benefit themselves at the expense of the workers. As a result, worker wages stagnate and even decline, and the fruits of economic growth go disproportionately to the elites. The operation of this wealth pump results in popular immiseration and discontent, elite overproduction and intraelite conflict, and gradual weakening of the state [my emphasis]. As these negative trends build up (which typically takes multiple decades), they eventually drive society into crisis. Exiting the crisis requires reversing these negative trends, which means, first and foremost, shutting down the wealth pump. But accomplishing this feat is not easy. The problem is that the wealth pump is extremely lucrative for the ruling class. Sociologists refer to this tendency of the ruling elites to use their power for selfish ends as the Iron Law of Oligarchy.* In short, power corrupts.
Peter Turchin, Cliodynamica, 26/5/25

An increase in union membership at two of Australia’s highest-profile technology companies, Canva and Atlassian, is small in terms of raw numbers but significant because of what it says about the disruptive impact artificial intelligence… That sense of workplace security in tech culture is quickly fraying as the sector confronts its Frankenstein moment: many of the workers who brought AI to life are now trying to survive before it turns on them… The experiences of Canva and Atlassian are symbolic of this cultural challenge and offer a broader lesson to other industries about the need to manage workers’ demands for transparency with the rollout of AI which is essential to a vibrant and productive private sector.
AFR editorial, 26/5/25

Is it realistic to assume humans will still be employable five or ten years from now? If not, why continue the charade? Many of our current decisions depend significantly on our expectations about the future—whether human labor becomes obsolete, or humanity itself faces extinction or radical transformation. Norms and conventions matter far less when we anticipate that the fundamental change is just around the corner.
James D. Miller, Aporia Magazine, 28/5/25

it’s possible that, by 2028, many of the issues that are crucial today will be swamped by concern about AI and its impacts on the world.
Matthew Yglesias, 28/5/25

Soon, ChatGPT will intuit what I want to vent about in my weekly post, generate the appropriate copy and image, and upload it to Substack. At that point, I’ll enjoy both utter frictionlessness and total redundancy.

For the time being, I need to wait for inspiration to strike. This week, there’s been plenty of that thanks to the first high-profile stirrings of Symbolic Analyst labour unrest in Australia. Or maybe Symbolic Capitalist unionism is a more attention-grabbing term? Anyhoo, a while back, Canva asked its staff to get with the AI tools program while reassuring them that they wouldn’t be training up their replacements by doing so.

Anyone who’s worked for a soulless megacorp for over five minutes has already figured out how this story ends. Canva laid off 10 of its 12 technical writers in early April. I’ll let the AFR, a publication not known for its pro-working man sympathies, take it from here:

Canva told the majority of its technical writing team on Friday that their services were no longer required. It is the first known redundancy round by the fast-growing technology group – ever…

Employees who spoke to The Australian Financial Review, under the condition of anonymity so they could speak freely, claimed the technical writers diligently followed the direction to use “any and every” generative AI tool in their work. Staff members said they had been assured by the company that the use of AI wouldn’t lead to a loss of jobs...

Canva sources said the redundancies had put employees on edge about other potential cuts the company could make as the accuracy of generative artificial intelligence tools such as OpenAI’s GPT and Anthropic’s Claude improve.

I’d argue Canva’s extant employees, and the rest of the knowledge-worker class, should be on edge about other cuts that will come as soon as the AI tools are sufficiently mature. As noted in recent posts, AI tools are already good enough for employers to now widely use them instead of human content creators, coders, grads and interns.

Shortly after reading the AFR’s dispatch about a “surge in union membership at the country’s two highest-profile technology companies, Canva and Atlassian”, I happened upon a Substack post by my fellow AI-obsessed Substacker, Brennan McDonald. The post is entitled Why Every Business has 5 Years to Achieve 90% Operating Model Compression. Here’s the money quote:

Forget about 10% FTE reductions as a playbook - this will be the era of 90%+ headcount reductions and customers telling their technology vendor they need to reduce next month’s invoice by 90% or they’ll rebuild their product internally.

Operating model compression flips the “10x developer” idea on its head - if you increase productivity 10x you only need 10% of the original resource footprint. That’s a 90% compression ratio.

Whether or not the compression ratio reaches 90 per cent, businesses will have no option other than to cut workforces to the bone ASAP. In the unlikely event that a business’s owners are kind-hearted enough to hesitate in reaching for the butcher’s knife, that business will be abruptly driven out of the market by its automated competitors. And that means a lot of soon-to-be former wage slaves will soon need to find some way to keep body and soul together.

As Yglesias notes, this will soon be the most salient political issue in technologically advanced, first-world societies. All the other problems – not least the rarefied post-material ones many tech workers got so worked up about circa 2020 – will fade into insignificance.

Digital-age wealth distribution
So, how will the benefits of AI-driven mass automation be distributed?

Will we continue with the neoliberal arrangement of the last half century and allow those at the top to capture the lion’s share of the exponential productivity gains soon to occur? Or could we return to a Keynesian dispensation, where wealth is shared more equitably?

The peasants are revolting
Several high-profile public intellectuals in both the legacy and independent media have recently taken it upon themselves to explain how horrific things were back in the Keynesian era. And how much better things are – most especially for Joe and Joanne Sixpack – after nearly half a century of goods, labour and capital flowing freely and wealth creators being appropriately venerated.

Here’s the concluding flourish from a recent New York Times piece by David Brooks:

The story the populists tell about globalization and neoliberalism is a gross distortion that leads to all sorts of terrible conclusions. America has many pathologies that drive the distemper of our times, but — at least until the populists gained power — economic decline was not among them.

Brooks magnanimously concedes there’s a housing, education and health care affordability crisis in the US, but compares putatively cashed-up US workers to European ones. That is, the ones with lower salaries but access to free or heavily subsidised housing, education and health care (and less Hobbesian existences overall).

If I wanted to argue the toss with the Brooks bros, I might mention the great decoupling of wage rises from productivity gains. Or maybe refer to the long-term decline in the share of GDP going to Labour and commensurate increases in the share going to Capital. Or perchance point to how C-suite compensation has grown vastly more lavish in recent decades.

But I won’t do that. I’ll simply point out that the average Anglosphere voter hews much closer to Turchin’s worldview than Brooks’. The common man and woman believe elites have been pumping that wealth pump a little too enthusiastically of late. Indeed, the cris de coeurs from Brooks, Cofnas, Smith and Yglesias are a tacit admission of how common terrible and distorted conclusions about globalisation and neoliberalism have become.

All the magic bullet solutions and why they won’t work
Neither Capital nor Labour – or perhaps that should be Post-Labour? New Labour, maybe? – is yet committed to an extremist position.

The Bannonite populists aren’t arguing that tech company staff should seize the digital means of production and start guillotining founders and their VC backers.

Elites aren’t ruling out redistributive measures like UBI. Many heavy hitters, including Sam Altman and Elon Musk, have long argued such measures will be required. Political gadfly and recovering tech bro Andrew Yang made UBI the centrepiece of his centrist US 2020 presidential campaign.

I expect an initial stage of happy talk with both sides expressing great hope of reaching a Grand Settlement. Expect to see a lot of references to the Treaty of Detroit, New Deal and, here in Australia, the Harvester Judgement.

If a grand settlement, or even just a tolerable one, is reached, nobody will be happier than I.

But given recent history – in fact, given all of history barring a few rare blips, such as the post-war decades – you’d have to expect those at the top to push hard for a distribution of wealth and power that’s more techno-feudalist than Keynesian.

What happens next?
Let’s war game how the final, existential battle between Labour and Capital might play out.

Stage one: White-collar workers – the type who previously assumed the nice lady from HR would always be there to advocate for them, even against the company paying her wages – will suddenly realise there is, potentially, power in a union.
Likely outcome: While I’m all for the workers uniting, it’s a little late in the day, on two counts. First, organised labour is a shadow of what it used to be. Not least due to the lack of enthusiasm private-sector, white-collar workers – a few modestly remunerated exceptions aside – have hitherto shown for unionism. Second, and I speak as a one-time print journalist and MEAA member, there’s not much a union can do for you once your employer has decided to get rid of you.

Stage two: Everybody decides retraining is the answer, but it soon becomes apparent it’s not.
Likely outcome: If a society has, say, a sudden surfeit of knowledge workers but a chronic dearth of tradies, why not just beat swords into ploughshares? It’s an enticing and not entirely wrongheaded idea, but retraining schemes haven’t proved a great success in the past. In the same way few middle-aged factory workers could learn to code, I suspect few middle-aged coders will transition seamlessly to becoming data centre maintenance men.

There’s also the issue of what those retrained knowledge workers will do when AI-equipped robots come for the roles that disembodied AI solutions can’t currently handle. This will probably happen within a decade, possibly within the next five years.

Stage three: Along with lots of money being pumped into retraining schemes, there will be many other piecemeal attempts to ameliorate the effects of growing unemployment. Standard working hours will be reduced, in the first instance, by moving to a four-day working week. The newly unemployed will be endlessly encouraged to become entrepreneurs and launch businesses. There will be a lot of talk about the importance of agency. (Many will desperately cling to the belief they can outrun the effects of mass automation if they can become sufficiently agentic.)
Likely outcome: A shorter working week means more work to go around, at least for a while. Some erstwhile junior employees will become successful entrepreneurs. A proportion of the population will be able to demonstrate more agency and maybe even reap some rewards for doing so. But as with retraining schemes, none of these measures will shift the dial much, especially over the medium to long term.

Stage four: Especially in a post-Luigi Mangione world, the alpha tech bros and their fellow captains of industry will solemnly declare they are fully committed to sharing the benefits of AI-driven productivity gains equitably. But while they’ll talk a good game, I predict unforeseen roadblocks and sudden reservations will materialise when it comes time for the wealth to start flowing from the fortunate few to the unlucky many.

Sure, a UBI sounds good, but is it affordable? Even if we can afford it, is it good for humans not to work? Won’t they become aimless and depressed if they don’t have a salt mine to slave away at?

Granted, a robot tax seems like a victimless crime, a painless plucking of feathers with no hissing from the unfeeling goose. But what if robot taxes slow down automation? Automation, even if it results in 90 per cent unemployment, is good because it means fatter profit margins more economic growth. And if there’s one eternal verity, it’s that a rising tide lifts all boats!

Of course, it might look like it is now feasible to fund entirely free education and healthcare, and maybe even housing and transportation. But will the punters be happy without a sense of struggle? What will people have to get up for in the morning if it’s not paying down their student debt while working towards saving the deposit on a grim studio flat in godforsaken outer suburb?

Stage five: This will kick off when it becomes undeniably apparent that the interests of a tiny but powerful – remember, they control the AI and robots – plutocratic class aren’t aligned with those of their fellow citizens. That’s when the real disruption will commence.
Likely outcome: Well, that’s the existential question, at least for the bottom eight or nine deciles of the income distribution, isn’t it?

But I’ve already gone on too long, so a deeper dive into Stage Five will have to wait for a future post. If you’ve got any theories about what happens when the redistribution rubber fails to hit the road, dear reader, please leave them in the comments.

*A theory put forward by sociologist Robert Michels that an elite class is unavoidable in any complex society. The problem isn’t the elite class, per se. It’s that the ruling class is, like all other classes – for the time being – made up of fallible humans.

Granted, they are usually high-achieving humans, at least in an occupational sense. But they are just as prone to mistaking their narrow self-interest for the greater social good as those with less impressive educational credentials. And as elites become increasingly self-interested, the societies they oversee become increasingly oligarchical.

**The dwindling band of neoliberal globalist true believers are beginning to sound suspiciously like those Che-Guevera-t-shirt-wearing undergrads who wave away the failures of central planning with the claim that ‘real socialism (free-market fundamentalism in this instance) – has never been tried’.