It seems that AI might not be stealing your job after all, and the reason for the constant news of job “reductions” is due to something more disingenuous.
For the past few years, all we hear from corporate leadership figures is AI, efficiency, productivity gains, and cost reduction. Whilst one of those points is not bs, the others almost always are simply said because corporate types love to repeat buzzwords their boss said.
The situation we are currently riding over follows something that is important to point out before we go any further. Back during the pandemic, we witnessed fluctuations in demand that gave the same corporations the balls to raise prices and go on a hiring spree that allowed them to take full advantage of demand spikes. Fast forward to today, that demand is long gone and these same companies are left with a sizeable workforce that is increasingly underutilised. The task they now face is how to fix this without looking like they over-hired in the past because they gauged the situation incorrectly, and instead frame it as a positive action.
Enter AI
OpenAI waltzed into our lives just as the pandemic was fizzling out of most people’s mouths, and with it came a gift for the corporations. They spoke of automation and agentic workers that would do the job of many humans all at a fraction of the cost.
Boardrooms got quite excited about this prospect because it basically answered their prayers. They now had a buzzword (office types love buzzwords just as minions love group gatherings) and it gave them a cover for what they were already considering.
Layoffs would have been a sign of overzealous hiring decisions, of being hasty or simply greedy; but now, it was a different message. The new story is that these layoffs are due to AI automation, increased productivity creating less need for actual staff numbers, etc etc. So instead of shareholder reactions to silly decisions by management, they had the perfect cover story to sell this as an efficiency move, as agility, as good management.
Just look at the news and all you see today is some stuffed suit talking about AI as if he or she had just met the messiah.
There is just one problem with this story.
The data does not agree.
The Data is a snitch
If AI were truly replacing workers at scale, output per worker should be skyrocketing; however, aggregate productivity growth remains “patchy” and pretty unexceptional.
What is being found at present (this might change over time) is that AI implementation is actually causing more work for adjacent colleagues. Often the data output needs to be checked for hallucinations or errors, which has been shown to frequently slow down throughput.
Where we have seen productivity gains, such as a 4.9% rise in Q3 2025, much of this is thought to be due to companies loading the same work onto fewer employees post-job cuts. AI does offer the potential to speed up work, it just does not seem to be the silver bullet we are being told it is, and the implementation is not quick or easy.
Evolution of the workplace
Similar to the introduction of the personal computer, AI is expected to destroy specific clerical tasks but create millions of new roles through “economies of scope”. We are already seeing jobs for “AI Literate” roles offering up to 28% higher salaries for candidates.
Furthermore, back in the 1980s, the car giants permanently downsized due to profits; whereas companies today still remain highly profitable, which suggests the motives for the changes are not in a similar vein.
There is also another bottleneck on the rollout of AI in the workplace and that is power. The physical infrastructure is struggling to keep up with projected demand and global power output must grow by up to 20% annually to just keep up.
On top of this, we have rising energy costs which could add further cost pressure on solutions to the point where the investment does not add up for now. Ultimately, energy-intensive compute power is at a premium, and to continue to expand from here will not come cheap or as easily as some would believe.
On the brighter side (I do not often get to talk about these!), AI is projected to actually create some roles that are focused on or related to its use. Estimates of up to 170 million new roles, such as prompt engineers and AI ethics officers, are just a couple that seem to be in the making.
So where does that leave us now? Well, AI is predicted to continue to automate mundane tasks and that will almost certainly be the case eventually. Some sectors will see huge improvements in efficiency once that is realised and at a sufficiently mature level; think HR, for example. However, what this could lead to, and let’s hope it does, is the releasing of workers into more creative roles where the human provides a more high-impact input that the AI companion assists with.
The future is never as simple as we sometimes like to believe; it is not A or B, and perhaps we should always try to keep that in mind. What is clear, however, is that corporate leadership has certainly combined their new favourite buzzword with a problem they needed to resolve, and we are being oversold on the AI solution.
Does this mean AI will not take jobs? No, that remains unchanged. However, it is not doing so today in the numbers we are being sold; the situation is a combination of some jobs being tested out with AI and some jobs simply being chopped and blamed on AI efficiency gains.
The problem for the companies dressing up the truth is that sooner or later they will need to come up with an even better tale to explain the lack of real productivity gains.
As always, time will tell all.
Damn, this hits hard. Been saying this for months, AI is just the new restructuring excuse.
the numbers don’t lie.