Last week, Microsoft released a report listing the top 40 jobs most at risk from AI — and the 40 least affected. No big surprises: roles like telemarketers and translators are near the top of the risk list, while hands-on jobs like nursing assistants and embalmers are much safer. But a new report from Harvard Business School’s Christopher Stanton suggests we may be underestimating just how quickly AI could disrupt even more white-collar work — and how little we might be able to do to slow it down.
Stanton, who researches how AI is changing the workplace, says the overlap between what AI can already do and what many office workers currently handle is bigger than we think. “About 35% of tasks in labor market data can be handled by AI today,” he told The Harvard Gazette. In other words, AI tools that are already available could potentially take over about one-third of the tasks in a typical office job.
Whether businesses actually will use AI to replace those tasks is still uncertain — but the capability is already there.
Still, Stanton does offer a more optimistic take too. Using his own job as a professor as an example, he says AI could take on 20 to 30 percent of a professor’s workload, like grading or scheduling. That would free up more time for things AI can’t do — like mentoring students, doing in-depth research, or engaging in meaningful classroom discussions. Ideally, AI wouldn’t eliminate jobs, but would help people focus on higher-value work.
But here’s the catch: the speed at which AI is rolling out might not leave enough time for that ideal future to take shape. Stanton warns that AI is spreading “probably faster than any other technology in history,” and that could upend the job market in ways we’re not prepared for. In particular, he worries about the impact on middle-class workers, who could find themselves replaced and with few options for new employment.
He’s not confident that politicians will be able to intervene effectively either. Outside of tax breaks or subsidies, he believes most policy efforts will fall short — mainly because companies that don’t embrace AI could quickly be outcompeted by more agile, AI-powered rivals with lower labor costs.
This tension — between AI’s promise and its risks — is playing out in real time at companies like Amazon. CEO Andy Jassy recently found himself in hot water after a poorly handled internal message about AI-led job cuts triggered a strong backlash from employees. Some even joked that senior executives’ roles should be next in line for AI replacement.
Jassy later struck a more careful tone during Amazon’s earnings call, saying AI will “very substantially change the way we work,” but that it will mostly help employees by removing tedious, repetitive tasks. The idea is that AI will make jobs more “enjoyable.”
That echoes recent research by HP, which found that 70% of workers who use AI say it makes their jobs easier — with reduced stress and better productivity.
Still, the reality is more nuanced. AI will take some jobs. But for those who remain, their day-to-day work might actually improve. It could become more focused, more strategic, and less bogged down by repetitive busywork.
So what should business leaders take away from all this?
If you’re introducing AI into your organization, how you talk about it matters just as much as how you use it. Be honest. Make it clear that AI is there to help, not to double everyone’s workload. And most importantly, don’t forget that behind every “productivity gain” is a person trying to make sense of what the future looks like.