Sometimes a historical trend is summed up by a single, dramatic event. Jack Dorsey announced that Block is to slash 40% of its workforce, around 4,000 people, in a move aimed at making the payments company he founded that operates Square, Cash App and Tidal, “move faster with smaller, highly talented teams using AI to automate more work.” What’s more, he predicts that most companies will do the same in less than a year. It’s not just the number: it’s the explicit normalization that AI is now about substituting people.
Most telling, however, has been the markets’ reaction: Block’s share price soared on the news. And that response, rewarding job losses, takes the debate on AI into an uncomfortable place. When financial incentives are aligned with sacking people, technology ceases to be a promise and becomes doctrine.
Some might say Block was overstaffed, that Dorsey has always had an “idiosyncratic” management style, and that AI is as good an alibi as any for a restructuring that would otherwise seem like butchery. Regardless, the damage has been done: the message is that a CEO can point to AI as the reason for the cut, and be rewarded for it. That precedent will send shivers down the spines of millions of white-collar workers who, until recently, felt relatively safe in their jobs.
And they’re right to be afraid. For years, automation has been sold as a way to replace repetitive, routine tasks, the implicit assumption being that the lowest-skilled will be hit hardest. The current generation of AI tears that narrative up. “Keyboard work,” from reporting to analysis, presentations, legal review, customer service, marketing, human resources, and more, is alarmingly compatible with what generative models know how to do. And in that compatibility there is a profound cultural change: many companies are no longer thinking about how to use AI to get their employees to do more, but how many they do without.
Harvard Business Review reports that some companies have jumped the gun, laying off staff on the basis of AI’s potential, creating a domino effect of hiring freezes, especially of graduates, thus replacing important first steps along career paths with premature automation. The IMF recently focused on that blind spot: entry-level jobs are more exposed and there is already emerging evidence (in the United States) that the adoption of generative AI reduces entry-level hiring when tasks are automatable. What is worrying is not only the employment that is destroyed, but the ladder that is removed: if the lower rungs disappear, how is experience acquired?
Which isn’t to say that those higher up shouldn’t be worried: international organizations have been warning for some time that senior profiles (managers, business professionals, engineers, scientists, etc.) are among the most exposed to advances in AI. The OECD pointed this out as early as 2023: exposure is likely to grow and spread as these tools are integrated into production processes. Translated: if you think this is about editors and customer service staff”, spoiler alert: wait for the next episode.
The ILO, refining its metrics with task-level data, is even more explicit: 25% of the workforce is employed in occupations with some degree of exposure to generative AI: clerical occupations appear to be the most exposed, and in addition, the impact is markedly unequal by gender and by the income level of their country. In other words, it is not only a question of the “future of work”, but the distribution of damage and power. And while the ILO correctly says that a transformation of tasks rather than outright job elimination is more likely, nuance matters little when business incentives turn transformation into downsizing.
Added to all this is the macro, almost reassuring discourse of the major reports: the World Economic Forum speaks of millions of jobs displaced and millions of new roles created, resulting in a positive net balance. Perfect… Except for one detail: “new roles” do not magically appear in the same places, with the same salaries, or for the same people. The issue is not whether there will be work in aggregate terms, but who pays the cost of the transition and who captures the productivity benefit. If AI is used to cut payrolls, the progress narrative loses any credibility.
The scary part is that this fear is spreading precisely where there used to be psychological safety: the office, the career, the promotion by merit, specialization. The white-collar worker begins to discover that their value was not “being smart,” but being scarce; and that if intelligence is packaged, priced and scaled, they’re no longer so scarce. Suddenly, salary negotiation is less like “what you contribute” and more like “what it costs to automate”. And when a company announces without a shred of embarrassment that a “significantly smaller” team can do “more and better” thanks to its tools, the message to those white-collar workers is simple: you are being replaced by AI.
The next step seems inevitable: what happens in the digital world will spread to the physical world as soon as robotics is no longer an expensive promise and becomes a commodity. When relatively cheap robots can perform tasks with sufficient reliability in logistics, hospitality, healthcare or light industry, blue collar workers will be next. The difference is that, in manual work, the pace is set by hardware, while in cognitive work, it is set by software. And software has no patience.
The important question is no longer whether AI will destroy jobs, but what we do when society accepts this and the markets reward it. If we really want AI to create shared prosperity, we must align incentives: worker participation in productivity increases, collective bargaining adapted to automation, transparency about what is automated and why, along with public policies that do not limit themselves to repeating “reskilling” like a mantra, and instead assume the problem of power behind it. The alternative is simple to imagine: an economy in which AI makes work cheaper by eliminating the worker.
Here’s one last irony: if the promise is that AI drives demand, efficiency and growth, what happens when the model to gain efficiency means laying off a growing part of those who consume? AI as an economic engine has an obvious limit if it becomes, above all, a wage-cutting machine. And that limit is not technological: it is social.
—
This post was previously published on Enrique Dans’ blog.
***
If you believe in the work we are doing here at The Good Men Project, please join us as a Premium Member today.
All Premium Members get to view The Good Men Project with NO ADS.
Need more info? A complete list of benefits is here.
Photo credit: iStock
