Reporting archive
Evidence on AI and work
Reporting archive

JPMorgan is graduating from chatbots to long‑running, permissioned agents—complete with audit trails—redesigning white‑collar work before it redesigns headcount.

Britain didn’t just talk AI—it shipped a 24/7 public job coach, and the next three months will reveal whether it turns layoff into interview time, not rhetoric.

Billions are flowing into AI, but the only obvious hiring is pouring concrete—because until costs fall and governance settles, “adoption” without real integration just adds steps, not pink slips.

The UK just turned “train and pray” into paid guarantees, rebuilding the missing first rung of the AI economy with contracts, not certificates.

A 4.95% vote at the nation’s largest employer signals investors prefer AI speed over standardized worker-impact metrics—and gives boards cover to keep the numbers in the dark.

In May, U.S. employers didn’t just talk about AI—they checked it on the layoff form, with 40% of cuts carrying the label.

BCG’s latest study shows AI has turned many roles from doing to directing—creating newfound hours that mostly leak away without a bold workflow rewrite.

When a $2.2T bank lets AI write first and makes humans prove what survives, the winners are editors, not typists.

While peers freeze hiring, Cognizant is onboarding 20,000 grads, formalizing AI-operator roles, and replacing token tallies with outcome guarantees.

A single payroll series just hijacked the AI layoff narrative, pointing to an implementation economy that’s hiring faster than automation can subtract.