End of Days – the Future of (less) Work with AI

In June 2020, ChatGPT 3 stormed into the public consciousness. We were right in the middle of the first wave of COVID-19, so needless to say many had a lot of time on their hands. We took notice, and even though it was (and is still) not perfect, people became aware of the potential that “true” AI could have in all aspects of our lives. We started to think, and dream, of a future where AI did all the work, freeing us from labour.

So what does AI mean for the future of work, and what does the actual “end state” look like? Many see a positive utopia, some sit in the middle and others… have their doubts. While I do already enjoy many of the benefits that AI brings, I’d probably put myself in the last camp. I am wary that AI is somewhat of an existential threat to society in general, and there is plenty of literature on how it is dumbing down our society, causing global instability, and destroying our environment. Can it really transform the Future of Work in a positive, meaningful, and sustainable way?

Let’s quickly look at some bullet points in favour of AI:

• It is a time saver for many menial tasks across many different fields

• It is much more flexible than previous attempts at cost reduction such as outsourcing or robotics

• It is largely accessible (at least right now) as almost anyone with access to the internet can utilise it in some way

• It is cheaper than humans for an increasingly impressive array of tasks

  
  

With every generation of AI (almost one a year over the last 5 years) we have seen capability skyrocket. “Accessible AI” (e.g. ChatGPT) has gone from a bit of a joke (how many fingers do humans have again?) to being able to code, assist in medical diagnosis, draw Miyazaki style images, create convincing videos and audio, research online, summarise, manipulate data and the list continues to grow. There are opportunities in almost every industry.

The existential threat of AI

As economies have matured they have increasingly moved away from manufacturing to knowledge industries. Not only has manufacturing capability been lost, but salaries and cost of living have risen to the point where it is not feasible to manufacture domestically. Just ask anyone in the garment industry in Australia - lots of designers, very few actually sewing. I once asked a local dress shop if they could make me up some Taekwondo uniforms to sell at my dojang. Her response? “The cost to make them would be more than you could sell them for”.

Over the last 30 years, as manufacturing has continued to decline, “Knowledge workers” have been where the jobs and wages growth has been. We pushed our kids towards STEM, and away from trades. We’ve now got plenty of software engineers, but we cannot produce enough chefs, plumbers, hair dressers, electricians, or aged and disability carers.

The rug pull is that AI impacts knowledge jobs harder and faster than any other type of work. So we have pushed our society into a specific type of work, and then introduced a tool that specialises in replacing that specific type of work…

We’re already seeing the impacts. Tech industry layoffs tracker Layoffs.fyi reports from March 2024 to June 2025 there have been 160,061 layoffs from 535 companies in the US,  compared with just 14,230 in the US during 2018. As I write this piece today, the first headline I read was about Microsoft cutting 9000 staff (4% of its workforce) in this year’s second round of redundancies.

As AI capability matures, it will encroach on more and more knowledge workers.

What does this mean for workforce planning?

Already, automation has removed many front-line jobs in the hospitality and retail industries. McDonald's used to need a dozen cooks to make their burgers - now, a handful simply assemble their components that are automatically cooked. Where there used to be a big row of manned check-outs at Woolworths, there’s now a self-checkout section with a couple of human cashiers. 

These “starter jobs”, mainly held by teenagers looking to build some experience, are important, because they help kids learn what it is like to be employed, and how to interact with others in the workplace. They also provide a pathway to management and even corporate function for those who don’t wish to, or are not compatible with, higher education.

So if you can’t get a job flipping burgers at McDonalds anymore, where do you turn? An internship, or other entry-level office role. These roles often provide the same sort of “starter job” experience for the corporate world. But just as automation has come for the entry-level manual labour roles, AI (in combination with off-shoring) has come for the office starter jobs.

In July 2025 it was reported that the UK’s Big Four accountancy firms had slashed graduate and school leaver intake over the past 2 years by up to 33%, citing AI efficiencies and a doubling down on outsourcing.

Even the countries that these outsourced workers come from are not immune. The Head of TATA Consulting Services gave an interview in 2024 saying that AI could kill call centres because theoretically the tech should be able to address the issue before the customer needs to make calls in the first place.

China’s youth unemployment rate reached a high of 21.3% in June 2023, and then they stopped publishing the figure. In the UK, youth unemployment is 14.3% and like Australia (9.47%) the trend is upwards. Young people are already finding it tough to break into the job market, and AI is only going to make it harder.

The compounded future impact

So we can see why replacing the new generation of youth in the workforce with AI and automation is bad for the kids specifically, but it is actually creating a ticking time-bomb for the company as well. You see, people age, and as your senior staff age out and retire, who is going to replace them if you haven’t been hiring enough new blood?

We have already seen this happen in the past on a mico-scale. Every time there is an economic downturn and graduate hiring within Professional Services is reduced, we see a shortage of managers approximately 3-4 years later. Now imagine this on a global scale across all industries.

This generational shift, plus the reduction in entry level jobs, will mean that those who hold the keys to the AI castles (the “humans in the loop”) will command premium salaries as fewer and fewer people will have built up the relevant experience due to the decimation of entry level roles. It’s a real downward spiral.

Over the next 5ish years, what I expect to see is more concentration of wealth and power. Big Tech owns the infrastructure and they are the ones who will make the most money. Uber is a great example of this - the restaurant owners and the drivers are not the ones getting rich. It’s the few at the top who control the software, who will demand more and more money in order to keep the system that they have effectively trapped us all in running, all while entry level jobs dry up and the middle management pipeline collapses.

Where does this leave us?

Universal Basic Income (UBI) is often the “endgame” that AI proponents speak of in a post-work society. “No-one will need to work because of AI, and everyone will simply get UBI and be free to do as they please.” But I really don’t think we’ll go down that path, because in my opinion UBI is simply economically unsustainable.

During the COVID-19 pandemic, we saw some countries flirt with UBI-esque policies. Australia’s infamous Jobkeeper program, which essentially paid people to stay home from their jobs, cost us $154 billion, and cleared out the government coffers so substantially that when the second wave of the pandemic hit (arguably more financially and mentally straining than the first), we simply had no money left. There were plenty of other UBI experiments around the world that have all failed - not to mention some more of the large-scale socialist systems like the Soviet Union (and we all know how that turned out).

In addition, research has shown that UBI robs people of the opportunity for people to find meaningful and fulfilling work. Governing.com’s Erin Norman writes “Working Americans seem to agree that employment provides not just cash but other intangible benefits. The State Policy Network poll shows that 3 out of 5 workers picked their current job because they wanted to help people, while 3 out of 4 believed that in their work they were currently helping people. Over two-thirds found their work personally fulfilling.” 

So what happens when we don’t have income?

Ivan 'Harry' Harrison has a great post about the unintended consequences of mass layoffs driven by automation. It introduces the concept of the "Automation Doom Loop," where job cuts lead to reduced consumer spending, ultimately harming the very companies implementing these layoffs.

Our global economies rely on people spending money. People can’t spend money if they aren’t earning money. The Automation Doom Loop is the bleaker side of the UBI coin, and unfortunately, the more likely path I see ahead for humanity.

What can we do about it?

The Godfather of AI, Geoffrey Hinton, has a blunt recommendation: “learn to be a plumber”. We’ve come full circle, and should rethink what skills are going to be useful in the future - we should value deep thinking, critical analysis over believing what the AI spits out and physically productive outputs over presentation decks and reports. Importantly we should stop thinking about “jobs” and start thinking about “tasks”.  As for me? Well, I’ve already hedged my bets. At TalentTech I translate between people and tech plus I operate and teach at my Taekwondo dojang.