60% of Employees Fear AI Is Eroding Their Skills, IBM Study Finds

Sixty percent of employees believe their skills are eroding as artificial intelligence reshapes daily work, according to a new global IBM study that surveyed 1,500 Chief Human Resource Officers and 8,800 workers across multiple sectors. Critical thinking tops the list of competencies employees say are slipping.

What do HR leaders and employees disagree on?

The research exposes a sharp mismatch between what HR chiefs say workers need and what workers themselves prioritize. Seventy-one percent of CHROs identified supervising and validating AI outputs as the most essential future skill. Only 29% of employees agreed that judgment in decision-making carries weight in the AI era.

Nickie LaMoreaux, Senior Vice President and Chief Human Resources Officer at IBM, framed the shift this way: "AI is changing not only how work gets done, but where people can contribute the greatest value."

The stakes are concrete. Companies that redesigned workflows into human-led, AI-assisted, or AI-executed categories reported a 20% improvement in quality and an 18% reduction in risk, the study found.

How deep is the skills concern?

Among employees who fear losing capabilities, 75% said AI has already begun to diminish their competencies. That gap between worry and lived experience signals an urgent retraining problem for employers.

Yet 46% of organizations did not involve their CHROs in shaping AI strategy. For smaller firms, where HR duties often fall on founders, that omission can stall adoption before it starts.

Who is accountable when AI fails?

Nearly half of employees said they feel personally accountable when AI goes wrong. That pressure discourages workers from challenging or overriding AI outputs, creating a culture of caution rather than critical feedback.

LaMoreaux pointed to a remedy. Organizations with clear human oversight, she said, see "a dramatic increase in employee confidence when interacting with AI." The implication: accountability structures, not algorithms, determine whether employees trust the systems they use.

What is the "invisible work" problem?

Eighty percent of CHROs said AI adoption generates hidden labor, from validating recommendations to handling exceptions that never appear on a project plan. The workload is real even when the tasks go uncounted.

This invisible load is particularly risky for small operators, where one staffer may absorb the hidden validation work without recognition or training support.

Why is HR itself slow to adopt AI?

Seventy-two percent of organizations reported limited or no AI adoption inside HR functions. That lag leaves recruitment, performance tracking, and employee management stuck in manual workflows even as the rest of the business automates.

The gap creates an opening. Small companies that build AI into HR early can align hiring criteria with the new skills CHROs say matter most, rather than retrofitting the function later.

What should small business owners do next?

The IBM data points to a clear playbook: involve HR in every AI decision, classify tasks as human-led, AI-assisted, or AI-executed, and fund continuous learning to rebuild the critical thinking employees say is fading. The firms that treat oversight and upskilling as core operating expenses, not optional perks, will be the ones that capture the 20% quality gains the study already documents. As LaMoreaux's team frames it, the question is no longer whether AI will change the workforce, but whether companies will redesign roles before their people redesign their careers.