Trending Articles
Trending Topics
Trending
AI is coming for jobs – we’ve been hearing that for a long while. No-one disputes its impact, but how many jobs and which jobs are big questions – and we’re starting to see signs that early warnings of a job-loss apocalypse were overblown.
Take Forrester’s latest report. This predicts that 6% of overall US jobs will be lost by 2030 because of AI and automation, so more than 10 million roles in that country alone.
It’s a big number, but that prediction is a lot more reasoned than what the doom-mongers at Oxford Martin were saying back in 2013, when they predicted half of all jobs would be automated. (That’s not quite what they said, but that’s how the headlines reported it – I’ll get to that in a bit.)
Now, anyone who has seen a robot stumble through conference halls or tried to make AI answer a useful question may not be surprised to hear that most jobs are safe from being wholly replaced by AI or automation in the near future. Nurses may get smart notetaking apps to do their paperwork and exoskeletons to help lift patients, but that just lets them get on with the actual work of nursing.
Forrester’s measured prediction hits back at the impression that AI is causing widespread unemployment in the here and now.
In a blog post, Principal Analyst JP Gownder noted that clients come to his consultancy saying that a fifth of staff are being laid off due to AI, begging for advice.
“When we ask if they have a mature, vetted AI app ready to fill in those jobs, nine out of ten times the answer is no – and they haven’t even started,” Gownder notes. “So, most of the layoffs are financially-driven and AI is just the scapegoat, at least today.”
He points to a long line of predictions about AI and job losses. For example, back in 2016 famed AI scientist Geoffrey Hinton predicted that we should stop training radiologists as it’s “just completely obvious” that deep learning will replace them within five years.
Hinton is a smart man, but he was very wrong. As Gownder points out, the radiology staff at the famed Mayo Clinic has grown by 55% since then.
And that’s for sensible tools like deep learning for computer vision for medical purposes, not for the flood of generative AI that’s been slipping slop into workflows since ChatGPT went public in 2023.
Anyway, Forrester does predict some job losses: 6.1% lost outright from AI by 2030. But it believes a fifth of jobs will be “strongly influenced” by AI in that same time frame. In short, we’ll still be working, but possibly with bots alongside us.
That’s not Forrester’s first prediction about AI and jobs, of course.
Back in 2016 – the era now known as pre-ChatGPT – Forrester forecast that 16% of US jobs will be replaced by AI or automation by 2025. It also predicted the technology would create the equivalent of 9% new roles, so it’s more like 7% lost overall.
By 2017, it was predicting a 17% job loss by 2027, with 10% new jobs, so it works out rather the same. By 2022, the consultancy was predicting the jobs created would nearly make up for the jobs lost, which at this point had slid to 7% by 2032.
Even post ChatGPT, Forrester towed the same line for its 2023 report: AI isn’t coming for your job, and if you’re in the 1.5% of roles where that does happen because of generative AI, there will likely be another one to replace it.
The point is this: Forrester originally said tens of millions jobs would be lost to AI starting in 2025. That didn’t happen. Depending on which report you want to read, it’s in the tens of thousands for July alone last year in the US.
At the same time, the US was hit by 1.1 million layoffs. Combining those two figures, and the rough maths suggest AI was responsible for perhaps 10% of total layoffs at best (or worst), rather than permanently removing that many jobs from the overall market.
Why are we all so worried AI is taking jobs?
Because the headlines are screaming it, and the AI developers love that kind of marketing to help fuel hype, so they’re not arguing.
Consider that Oxford Martin study that suggested 47% of jobs would be lost to AI. Those researchers weren’t talking about generative AI, which didn’t exist yet, instead totting up the number of jobs that had some repetitive aspects, which could one day be automated if a relevant technology was created. That’s… it.
As one of the researchers noted a few years after the hullabaloo: “Our estimates have often been taken to imply an employment apocalypse. Yet that is not what we intended or suggested. All we showed is that the potential scope of automation is vast…”
We’ve now reached the point where we’re losing jobs to AI, rather than the five years from now predictions. Pinpointing exactly how many jobs isn’t easy, though, not least because we can’t take companies’ claims at face value: can we really trust them when they list AI as the main reason for cuts?
After all, companies may be cutting jobs under the guise of AI simply to shore up budgets, as Gownder notes, or to invest in AI – not because they know the tech can do the job.
Plus, some companies – I’m looking at you, Klarna – have cut roles due to AI but quickly turned around and hired back human staff, though not necessarily the same ones or in the same numbers. Indeed, some companies see this as an opportunity to reboot offshoring: use AI for some tasks, and hire cheaper overseas workers for anything else not yet possible to automate.
So is your job safe? It depends what you do, of course, but AI is just one of very many reasons that you might lose your job this year or next. But as the good folks at Forrester show, one job that is secure is attempting to predict the impact of AI.
