Are we heading into a ‘Great Automation’?

It’s not just AI – all kinds of emerging tech is muscling out human workers

Photo by julien Tromeur on Unsplash

Two-minute video summary

Corporations teamed up with politicians of both parties to smash FDR’s New Deal. They replaced it with an order characterized by massive inequality, low wages, precarious service-based gigs, and Wall Street’s overweening control over the real economy. The ILA strike is a rare reminder of working people’s power to shut it all down. And it was exactly such a hard-line approach that created the conditions for the New Deal… The asset-rich had it so good over the past few decades—capturing the lion’s share of the upside from de-unionization, financialization, and offshoring, as wages stagnated for the bottom half—that they all but forgot what labor militancy can look and sound like.
Sohrab Ahmari, The Free Press, 3/10/24

The way Telstra boss Vicki Brady sees it, telcos are on the cusp of another big shift. Artificial intelligence is already starting to reshape her industry as much as other turning points such as the rollout of broadband; the move from analog to digital; or the arrival of the smartphone…

However, Brady is a realist: Like the evolution of any technology, it ultimately means there will be some functions that move to full automation and jobs will be lost. “All jobs are going to look different in five years’ time,” Brady says. “Our approach is to invest in our people, to leverage this so that it makes their jobs better, delivers better outcomes for our customers. But of course, as you deploy new technology, you’re always looking to be more efficient in the way you run your business as well.”
Eric Johnson, The Australian, 1/10/24

What our American friends call longshoremen (i.e. wharfies or dockers) are currently getting a taste of what “looking to be more efficient in the way you run your business” looks like in practice.

As you’re reading this, 45,000 International Longshoremen’s Association (ILA) members are out on strike. As is traditional, they want higher wages. But they mainly want to prevent the automation of cranes, gates, and container movements at US ports. Such automation has already occurred at many Asian and European ports.

These kind of ‘rage against the encroaching machines’ strikes are becoming increasingly common. Hollywood actors and writers downed tools last year in an (almost certainly futile) attempt to strictly regulate, if not ban, the use of AI.

Unsurprisingly, studio executives were delighted at the prospect of getting ChatGPT to churn out the script for the next Marvel movie and subbing in digital avatars for expensive, temperamental actors. An uneasy, and almost certainly temporary, showbiz truce was brokered in late 2023. (The bosses made some probably unenforceable commitments, and the creatives returned to work.) In all likelihood, a similarly face-saving compromise will eventually be brokered between the longshoremen and their employers.

But if I were advising the International Longshoreman Association, I’d tell them that while a bolshie union with the ability to inflict significant economic pain may be able to delay automation, there’s no standing in the way of progress. The best longshoremen can hope for over the longer term is what their European counterparts at ports such as Rotterdam negotiated. That is, some form of golden handshake.  

Are the frogs starting to notice the boiling water?
In nations where dockworkers’ unions are either non-existent or more pragmatic, docks have already been substantially automated. Dockworkers in the US, and Australia, may hold out for a bit longer, but there is only one possible end to this story.

For, as Noah Smith points out, the economic self-harm businesses and nations that drag the automation chain do is substantial and, in the long run, unsustainable: 

Banning automation is just a way of killing the goose that lays the golden eggs, ultimately hurting dockworkers. Meanwhile, it acts as a tax on the whole U.S. economy. If you didn’t like the inflation of 2021, you should want more efficient, high-capacity automated ports. The U.S. should emulate Rotterdam instead of retreating into self-defeating luddism.

Of course, it’s easy for Smith and I, as members of the professional-managerial class, to insist dock workers should take one for the team and graciously accept the automation of one of the few industries still providing well-remunerated work to blue-collar men.

But, for once, this isn’t a class issue.

We’re all dockworkers now – with a tsunami of automation bearing down upon us.

Remove the friction, remove the jobs
You had three options if you wanted to see a film when I was a boy. You could watch it at a cinema, borrow a giant (VHS or, for a time, Beta) cassette tape from a video store, or wait for it to be screened on a free-to-air TV channel.

In short, there used to be substantially more friction, to use the tech-industry terminology, involved in watching a film. But that friction was the reason cinemas, video stores and TV networks used to employ far more people than they do today.

Does that mean I want to ‘de-automate’ the viewing of films?   

Well, no.

Even if I did, the idea would be absurd. There’s no turning the clock back.

However, it should be recognised that Schumpterian creative destruction involves vaporising people’s livelihoods. 

Netflix employs around 13,000 people, mainly in the US and Canada. Even if you add in all the other people employed by the other streamers, it would still be a tiny fraction of the number of people who used to work at video stores, let alone video stores, cinemas and TV networks.

Most erstwhile video store owners and staff presumably went on to have perfectly happy lives. But there must also have been plenty of casualties when the industry collapsed. Small businesspeople who mortgaged the family home to buy a video store only to find revenues collapsing well before the loan was paid off. Aspy middle-aged film nerds not well suited to other forms of employment. And so on.

And there’s the rub. Greater efficiency is almost always a corporate euphemism for lower headcount. If you arrange your holidays online, that reduces the demand for travel agents. If you bank online, that translates into fewer bank tellers. Taking pictures on your phone rather than using a camera and film obviates the need for photo developers.

As a consumer, minimal friction and maximal efficiency are fantastic. Who doesn’t love effortless convenience, especially if it also results in cheaper prices?

But one person’s increased ease is another’s redundancy.

What if this time it really is different?
At this juncture in think pieces about automation, it’s the done thing to observe humans have been descending into ‘the robots are taking our jobs’ panics for centuries and that new technologies have invariably generated enough new jobs to replace the old ones.

This is entirely true. It’s also true that the issue facing Anglosphere nations in recent years has been having too few workers rather than too many.

Nonetheless, I suspect we are now on the cusp of an epochal shift that will see much of the work currently done by humans – on docks, in law firms, in shops, in restaurants, on farms, in insurance companies, at real estate agencies – automated away.

Not all of it will be automated away, at least not anytime soon. For instance, I don’t imagine tradesmen or nurses will have anything much to worry about for at least another decade.

However, technological unemployment is a real possibility unless you work in a role that a software program, AI agent, or industrial robot can’t readily take over. A 2023 World Economic Forum (WEF) report noted, “Employers anticipate a structural labour market churn of 23% of jobs in the next five years,” and that, “Of the 673 million jobs reflected in the dataset in this report, respondents expect structural job growth of 69 million jobs and a decline of 83 million jobs. This corresponds to a net decrease of 14 million jobs, or 2% of current employment.”

Given the WEF’s neoliberal enthusiasms, it’s incentivised to talk up automation but even it is predicting a decline of 14 million jobs. I’d predict the job loss between now and 2030 to be far more significant. And I’m not the Lone Ranger there.    

It’s early days, but I suspect workers fall into three broad categories. There are a substantial minority whose jobs will be impacted but not eliminated (anytime soon) by automation. There is a much smaller group of cognitively elite workers who will have their current role disappear but who will be able to retrain, probably as AI/ML specialists, data analysts, cybersecurity specialists, robotics engineers and so on. Then there’s everybody else – people likely to find themselves in the same situation video store owners and employees were in circa 2010.     

Could 2025 be the year the automation rubber hits the road?
We now appear to have gone through the initial two phases of the inevitable hype cycle – the ‘peak of inflated expectations’ and the ‘trough of disillusionment’.

(To be fair, a rapidly dwindling band of thinkers argue we are still in the unrealistic expectations phase and will eventually be plunged into despair when it turns out AI has been massively overhyped. If those contrarians are proven correct, please forget I ever wrote this blog post.)

But it appears we are headed into the slope of enlightenment/plateau of productivity phases in 2025. In concrete terms, that means tech companies that have poured billions into AI will start offering AI-based products to businesses and government departments that promise “greater efficiency” (i.e. significant headcount reductions).

Tech industry behemoths have made big bets on AI. The CEOs of these behemoths are now expected to start getting a decent return on that investment.
Fortuitously for the likes of Satya Nadella, boards are salivating at the prospect of slashing labour costs and are putting pressure on CEOs to put the AI-enabled automation pedal to the metal ASAP.