Rehiring after AI layoffs is not a contradiction: it is the sign that workforce cuts moved faster than learning

Few narratives have taken hold as forcefully as that of the 'AI layoff.' That is why it is worth pausing on the contrary signal picked up by WFAA, ABC's affiliate in Dallas-Fort Worth: a survey reportedly indicates that several companies are rehiring workers they had previously let go in cuts attributed to automation. Let us begin with methodological honesty: the accessible material is limited —it is a video piece without a full transcript—, so it is not possible to pin down the sample, the sectors or the figures without falling into speculation. What is clear is the direction of the movement.

Far from being a paradox, this swing fits with what some labor economists describe as a 'skills adjustment.' Organizations eliminate positions whose tasks are automatable and, after a transition period, discover that they need people with different profiles: capable of operating, supervising and improving those very systems. It is not that AI does not replace tasks; it is that, in doing so, it generates others that someone has to take on.

Agentic AI illustrates the point better than any other segment. An autonomous agent can execute complete workflows, but it rarely operates without a safety net: it requires human supervision, results auditing, instruction maintenance, integration with legacy systems and management of the errors it inevitably makes. Each of those responsibilities is a role that, in many cases, did not exist before automation. The work does not simply disappear; it shifts and changes in nature.

The moral, read with caution, is less dramatic and more demanding than the tale of total replacement. Those who rush cuts without a capabilities strategy run the risk of having to retrace their steps; those who pair AI adoption with reskilling plans start with an advantage. It is a good reminder that technology advances faster than the organizations that implement it.

For the sake of rigor, we insist on the essential: without access to the full breakdown of the survey it is not possible to assert magnitudes or generalize to the entire economy, and the recommendation is to go directly to the source to learn the methodology. Even so, the episode adds a necessary nuance to a debate too prone to extremes: between the apocalypse of employment and the denial of change, reality seems to be advancing along the more prosaic path of adaptation.