AI appears to be making workers faster, more efficient, and more productive on paper.
However, according to Mehdi Paryavi, CEO of the International Data Center Authority, which advises companies and governments on building the data centers that power AI, it may also be quietly eroding workers’ confidence in their job skills.
Paryavi told Buisness Insider that excessive and poorly designed AI use in workplaces is driving what he called a “quiet cognitive erosion” and “down-skilling.”
“There used to be a notion called ‘thinking outside the box,'” he said. “That notion will soon cease to exist when everyone draws on all their creativity, analytics, and innovation from a single box called AI.”
Confidence is the first thing to go
Paryavi believes the most immediate casualty of heavy AI reliance is self-belief.
“If you come to believe that AI writes better than you and thinks smarter than you, you will lose your own confidence in yourself,” he said.
Paryavi said the loss of confidence compounds quickly as workers begin deferring writing, analysis, and judgment to AI systems, gradually relying less on the skills they have built through years of reading, writing, learning, and observation.
“All of a sudden, you realize you are not good enough without this new tool, and day by day, you rely less on yourself and more on AI,” he said.
Research is beginning to reflect that pattern. A new report from the Work AI Institute, produced with researchers from universities including Notre Dame, Harvard, and UC Santa Barbara, found that AI is turning ordinary office workers into people who feel smarter and more productive while their underlying skills slowly erode.
Rebecca Hinds, head of the Work AI Institute, told Business Insider that AI creates an illusion of expertise, which is especially risky for early-career employees who still need to establish their foundations.
Speed isn’t the same as productivity
Much of the problem, Paryavi said, lies in how leaders define productivity.
AI’s biggest promise is speed — faster reports, faster launches, faster analysis. But faster doesn’t always mean better, Paryavi said.
While AI can generate professional-sounding output, Paryavi said it often lacks the depth that comes from years of hands-on expertise.
That loss of depth is already visible, according to Anastasia Berg, a philosophy professor at the University of California, Irvine, who has said that workers who rely heavily on AI risk rapid skill atrophy, especially junior employees who never fully learn how to think through problems independently.
How to avoid cognitive dependency
Paryavi isn’t opposed to AI. He said the risk comes from indiscriminate use.
Companies should tailor AI access by job function, he said, rather than rolling it out universally. Some roles may benefit heavily from AI support, while others should rely primarily on human judgment.
He also discussed the importance of human involvement at both ends of the workflow — leading creative thinking at the beginning and quality-checking AI output at the end.
“What’s critical to note is that you, the human you, must quality check AI, not the other way around,” he said.
AI may not eliminate jobs outright. But without deliberate limits, Paryavi said, it could quietly erode the confidence and thinking skills that careers are built on.
“How much technology do we really need, and how far are we willing to push the envelope? How much is enough?” he said.
