AI is often cited to justify layoffs but that masks how the labour market is actually distorted by it

Singapore-based Crypto.com said last week that it was cutting 12% of its workforce. Earlier, Atlassian and Block had cited AI adoption for job losses. But what’s lacking in these pronouncements is the evidence of how, exactly, AI is replacing workers.

Comprehensive data on whether Ai is destroying jobs, lifting productivity or reshuffling routine tasks remains patchy at best. And that vacuum is being filled with fear-mongering and market-friendly spin. No doubt, it is reshaping how people work, but for governments and business leaders to effectively react, far more data and transparency is needed.

In a world where engagement algorithms shape public speech, the loudest voices are rarely the most nuanced. Stories of white-collar bloodbaths have gone viral in recent weeks—and even moved markets —despite little hard evidence.

The narrative is potent. But the bigger danger may not be that AI is causing a ‘jobocalypse,’ but that these headlines obscure how AI tools are being used to erode the entry-level roles that train tomorrow’s workforce. This ‘AI washing’ becomes a distraction from the harder policy work required for periods of rapid technological change.

Investors are rewarding AI washing, but should they? Recasting pandemic over-hiring and cyclical belt-tightening as innovation and efficiency may help send share prices higher in the near term, but they don’t vouch for sound fundamentals or wise management.

The layoffs narrative does not neatly fit in Asia. In Japan, one survey found that nearly 30% of 246 listed companies were increasing their workforce after adopting AI.

An OECD report published last October argued AI-induced job losses may be less common in Japan than elsewhere due to chronic labour shortages driven by demographic decline and people’s tendency to stay at one company for long periods. Japanese workers, it found, are more likely to see AI as a source of new jobs than a destroyer of them.

A similar tension is emerging in South Korea. IMF researchers say that while about half of jobs are “exposed to AI,” the negative effects of an ageing population could be mitigated through AI adoption.

That isn’t an argument for complacency. Older workers, non-regular employees and entry-level staff are less likely to benefit from the shift, which makes training and workplace adoption programes all the more important. The real balancing act is how to use AI to ease labour shortages without allowing it to widen inequality or disrupt livelihoods. This is where Asia can show leadership.

The gap between what is known and what is being claimed is already too wide. Breathless predictions of how all white-collar work will be automated within 18 months aren’t analysis, but marketing. And they’re also bad for the technology itself. Public trust in AI is fragile. Tech leaders who want it widely adopted should stop selling every restructuring as proof that machines are turning humans obsolete.

Even if the overall scale of layoffs is being overstated, early signals suggest pain for entry-level workers. It may make short-term business sense to use a model for tasks once handed to an intern. But it’s a shortsighted bargain.

One of AI’s biggest limitations is still hallucinations. Human oversight remains essential for its use in businesses, hospitals and elsewhere. But people cannot check a machine’s output if they have no expertise themselves.

Companies risk hollowing out the apprenticeship layer where knowledge workers learn by doing. In my line of work, for example, an experienced editor can spot the cliches, repetitive phrases or dramatic but inconsistent metaphors that AI tools love to use. An engineer who has reviewed hundreds of design drawings can spot when an AI tool’s elegant solution will fail in the real world.

That should worry governments as much as employers—especially in Asian countries where a Gen Z job crisis has been brewing. Failing to invest in the next generation of talent will backfire. Large numbers of unemployed educated youth do not make for a stable society. Companies, universities and policymakers need to do more to protect training paths and junior roles.

Lawmakers trying to tackle AI’s impact on jobs should start by asking companies that cite AI as a reason for layoffs to disclose what that actually means: where the technology was utilized, what work changed, what productivity gains were measured and how many jobs were truly eliminated as a result. Only then can governments build sensible responses, from stronger social safety nets to targeted training and reskilling programmes.

The author is a Bloomberg Opinion columnist covering Asia tech.