The damage did not require AI to be good. It only required AI to be expensive.
82% of executives report decreased perceived value of human workers, despite 73% acknowledging AI underperformance and 69% spending more time monitoring AI outputs. Management appears to blame workers for AI's failure rather than acknowledging implementation errors, with 88% suspecting employees use AI performatively without adding real value.
- 82% of executives report decreased perceived value of human workers
- 73% of those same executives admit AI failed to meet expectations
- 69% now spend more time monitoring AI outputs rather than delegating
- 88% suspect employees use AI performatively without adding real value
- Survey covered 2,850 senior leaders across five countries
A G-P survey reveals 82% of executives perceive lower value in their employees due to AI adoption, even though 73% admit the technology failed to meet expectations, suggesting symbolic devaluation precedes actual job displacement.
On May 12th, G-P, a global employment services firm, released its third annual report on artificial intelligence in the workplace. The headline number was stark: 82 percent of executives surveyed admitted that AI had lowered the value they assign to their human workers. The survey reached 2,850 senior leaders—vice presidents and above—across the United States, Germany, Singapore, Australia, and France.
The easy reading of that statistic feels inevitable. Machines advance, humans become redundant, bosses notice. Progress marches forward, and workers fall behind. But the same report contains a second number that complicates that story entirely. Seventy-three percent of those same executives said AI failed to deliver what they expected. Sixty-nine percent now spend increasing amounts of time simply monitoring what the AI produces—a burden the report calls the hidden tax of adoption. The share of companies pursuing aggressive AI innovation dropped from 60 percent to 42 percent in a single year. And 44 percent expect the AI bubble to burst before year's end.
So the question shifts. If the tool disappointed, why did the worker lose value?
For three years, the dominant narrative placed the problem squarely on workers. They resist change. They use AI performatively, without genuine commitment. They fail to adapt. The G-P survey captures this suspicion: 88 percent of executives believe their employees simulate AI adoption without creating real value. But the numbers tell a different story. The executive who purchased expensive technology, who promised returns to the board, who now monitors outputs instead of delegating—that executive made the bet. The investment failed. And here lies the shift: instead of concluding "I failed to implement this correctly," much of management appears to have concluded "the person next to me is worth less."
Frustration with their own failed gamble transformed into a reassessment of someone else's worth.
It matters to be precise about what the report measured and what it did not. G-P did not count layoffs. It did not measure actual productivity. It measured perception—it asked executives how they view their workforce, and 82 percent said they view it with diminished value. That is not data about the labor market. It is data about how management thinks.
And that distinction carries weight. The mass replacement of workers that has been announced since 2023 has not yet occurred. Serious studies of employment and wages continue to show small or negligible aggregate effects. But the symbolic devaluation has already happened. The employee still sits at their desk, still receives a paycheck, still produces work. What changed is that the person who controls their future now sees them differently—and sees them differently because of a tool that did not even work as promised.
The damage did not require AI to be good. It only required AI to be expensive.
One detail shapes how to read this story. G-P is an employer of record firm—it helps companies hire workers across borders without opening subsidiaries. The company sells, among other things, its own AI-powered human resources agent. The report is, in part, a marketing document. That does not erase the data, but it demands careful reading. A company whose business depends on placing human talent across borders has every reason to emphasize that talent remains essential. And yet its own survey let the 82 percent escape. When an uncomfortable number appears even in the report of someone with incentive to soften it, it deserves serious attention.
The question most people ask is whether AI will replace them. The G-P report suggests that may be the wrong question, or at least a premature one. The actual question is whether an executive who approved an AI budget that failed to deliver will process that failure as a management error or as evidence that the people under them matter less. Eighty-two percent have already answered. And that answer does not depend on whether the technology improves. It depends on how management chooses to explain its own disappointment.
Citas Notables
The reemplazo masivo that has been announced since 2023 has not yet occurred, but the symbolic devaluation has already happened.— Analysis of G-P survey findings
La Conversación del Hearth Otra perspectiva de la historia
So the executives are saying AI didn't work, but they're blaming the workers anyway?
Not quite blaming—more like reframing. They spent money on a tool that disappointed them. Rather than own that as a bad investment or poor implementation, they're interpreting the disappointment as evidence that their workers are less valuable than they thought.
But that's backwards. If the AI failed, shouldn't that prove workers are still necessary?
You'd think so. But perception doesn't work that way. The executives made a bet. The bet lost. Now they need a story that makes sense of the loss. "The technology wasn't ready" is one story. "The people I hired aren't as good as I thought" is another. The survey shows which story they're choosing.
Has anyone actually lost their job because of this shift in perception?
Not according to the data in this report. No mass layoffs yet. But that's almost the unsettling part—the devaluation is happening before the replacement. Workers are still employed, still paid, but they're being seen as less valuable by the people who control their futures.
Why does that matter if nothing has changed yet?
Because perception shapes decisions. If a manager sees you as less valuable, that affects raises, promotions, which projects you're assigned to, whether you're kept during a downturn. The symbolic shift comes first. The material consequences follow.
And the executives know they're doing this?
The survey asked them directly, and 82 percent admitted it. So yes—they're aware enough to tell a researcher, even if they might not frame it to themselves as "I'm devaluing my workers because my technology purchase failed."