Read enough coverage of AI and jobs in a single week and you come away genuinely unsure whether the sky is falling or clearing. A Washington Examiner op-ed this week declared the "AI kills jobs" myth to be collapsing, pointing out that Geoffrey Hinton told the world to stop training radiologists in 2016 and that radiologists are now among the most sought-after doctors in America, facing their largest shortage on record. It cited a Ramp study of more than 21,000 companies finding that firms investing heavily in AI grew their headcount by 10.2% over two years, with entry-level hiring rising even faster. In the same news cycle, workers in Kochi were losing their jobs at a health-data firm where a process that once needed 100 medical coders now needs about 20 people to check the machine's output.
Both stories are true, which is what makes the debate so slippery. The headline numbers barely constrain each other. The World Economic Forum projects 92 million jobs displaced by 2030 and 170 million created, a net gain that depends entirely on whether the new roles land near the people who lost the old ones. Challenger Gray has counted more than 71,800 US layoffs that explicitly named AI since 2023. PwC, meanwhile, reports that job postings requiring AI skills grew nearly eight times faster than the overall market last year, carrying a wage premium of up to 62%. You can assemble a confident case for optimism or catastrophe using only verified figures, which should make everyone a little more humble about their favourite one.
The most useful reporting this week did something harder than pick a side: it questioned the premise. An analysis at Sify flagged a survey in which six in ten companies admitted they framed layoffs as AI-driven even when the real reason was financial. If that holds, then a large share of "the robots took my job" is really "the quarterly earnings call took my job," with AI serving as the more palatable explanation. That reframing matters, because it moves the villain from an unstoppable technology to a very stoppable management decision.
You can see the decision cut both ways in the same industry. Oracle shed tens of thousands of roles while pouring money into AI data centres. JD.com's founder, Richard Liu, looked at an automation programme deep enough to include unmanned warehouses and drone delivery, then promised not to fire a single frontline worker, signing up around 120 schools to retrain couriers into robot technicians. He was not being sentimental. PwC's data suggests the companies using AI to grow, rather than to cut, are the ones pulling ahead on productivity and even on hiring. Firing your workforce to save money increasingly looks like the strategically worse bet, not just the colder one.
Which is why the argument is quietly changing shape. The economist Nouriel Roubini warned this week that AI could eventually force governments toward universal basic income, or toward taking ownership stakes in the companies capturing the gains. A European Business Review essay made the same point more plainly: the real question was never whether AI destroys jobs, but who benefits from the wealth it creates. Productivity rising while inequality widens is not a hypothetical; it is most of the last forty years. The technology will keep improving either way. The part still genuinely undecided is who gets to keep the upside, and that has never been a question the data alone can answer.