The familiar question about AI and work is how many jobs it will destroy. A cluster of reporting this week suggests a more unsettling one is coming into focus: even for the people who keep their jobs, AI may be quietly draining their leverage. European Central Bank chief economist Philip Lane warned that the technology could weaken workers' bargaining power, putting downward pressure on wages and shrinking labour's share of income. That is not a story about empty desks. It is a story about who sets the terms at the desks that remain.
India offers an early, concrete look at the squeeze. White-collar hiring there grew just 1 percent year on year in September, and hiring of entry-level workers rose by the same thin margin, even as demand for AI and machine-learning skills jumped 20 percent. A Nomura analysis counted 83,100 AI-related hires against nearly 32,000 AI-linked layoffs, a reminder that creation and displacement are happening at once. But the new roles often demand different qualifications from the old ones, so a worker whose support job is automated cannot simply slide into an AI engineering seat. When openings in your field dry up, you lose the quiet power that comes from being able to walk, or to wave a rival offer at your boss.
Human Rights Watch argues the clearest preview of an algorithm-managed workforce is already running, in the gig economy. Its research across the United States and countries from Kenya to Pakistan found platform companies using opaque, frequently changing algorithms to set pay, ratings and access to work, with little way for workers to contest a bad score or a sudden deactivation. A delivery rider in Delhi described breaking his arm while rushing to hit a target, then paying for treatment himself. "We don't get anything from the company," he said. Firms insist these workers are independent contractors even as their systems direct the work, shifting cost and risk onto the people doing it.
There is, for once, a concrete counterweight. In June, governments, employers and unions at the UN's International Labour Conference adopted Convention No. 193 on decent work in the platform economy by a vote of 406 to 8, the first binding international standard to regulate algorithmic management. It requires ratifying governments to let workers request written explanations of significant automated decisions, to guarantee human review when pay or access is withheld, and to classify workers by how the job is actually done rather than by the label on a contract. HRW notes the stakes reach well beyond ride-hailing: algorithmic management has spread into domestic work, beauty services, logistics and health care.
You can see the anxiety playing out in the aisles of a Los Angeles supermarket, where inventory robots named Tally now roll past shoppers scanning shelves. Simbe, their maker, says more than 3,000 have been contracted and that they free staff for better work; the grocery union sees jobs and hours quietly draining away, and worries the machines could monitor employees too. Kroger, which runs the stores, said this week it will cut 3,000 corporate jobs over three years as it leans harder on AI. The through-line from a Delhi delivery rider to a Ralphs night shift is not robots seizing the future. It is a slow redistribution of power, and the open question is whether rules like Convention 193 arrive fast enough to matter.