Ford is rehiring hundreds of experienced engineers for quality-control work its automated systems could not handle, and IBM has said it will triple US entry-level hiring across all business units in 2026 after its own artificial-intelligence (AI) human-resources system failed the hardest 6% of requests1. Charles Poon, Ford's vice-president for vehicle hardware engineering, put it plainly: "Artificial intelligence is a fantastic tool, but it's only as good as the information you use to train it."
Orgvue first quantified the regret in March, when 55% of leaders who had cut for AI called the decision wrong, and Klarna had by then rehired customer-service agents after admitting it "went too far" on automation. The staffing firm Robert Half now adds fresh payroll evidence, reporting 32% of US hiring managers eliminated a role for AI, then rehired for the same or a similar one2. Ford and IBM turn that survey signal into hiring action at industrial scale, extending the overshoot this beat has tracked since ResumeBuilder found 59% of firms had overstated AI's role in their cuts, the pattern MIT Sloan's Paul Osterman described when he called AI attribution a cover story for pre-planned reductions.
Both reversals point to the same limit, the last-mile problem: automation clears the routine bulk cheaply but breaks on the judgment-heavy residual, and the cost of that failure, whether a vehicle recall or a mishandled hiring case, can exceed the wage bill it displaced. IBM's chief human-resources officer Nickle LaMoreaux framed the correction as pipeline defence, warning that cutting entry-level hiring now means "the well simply dries up" in three to five years3. None of the July evidence is payroll-hard at the level of an official series; the corroborating surveys are vendor-adjacent, and no national dataset yet confirms net AI-driven rehiring at scale.
