A new IBM study of 1,500 CHROs and 8,800 employees across 28 countries finds that HR chiefs and the people they manage have started to disagree about the one skill that matters most in the AI era: judgment. The gap is not academic. It is already showing up as blame, invisible work, and eroding confidence on both sides of the org chart.
The IBM Institute for Business Value, working with Oxford Economics, surveyed CHROs and senior workforce-strategy executives across 21 geographies and 23 industries between April and June 2026, alongside a separate poll of full-time employees. The headline finding is a split screen. Seventy-one percent of CHROs say the most essential workforce skill now is the ability to supervise, validate, and override what AI produces. Only 29% of employees rank judgment that highly. Meanwhile, 60% of employees say they are worried AI is eroding their own skills, with critical thinking named most often as the capability in decline.
Why this is a workforce problem, not a technology problem
The disagreement matters because it determines who gets blamed when AI gets something wrong. IBM’s data shows 43% of employees report that when an AI system makes a mistake, the responsibility lands on them personally. At the same time, 41% of CHROs believe employees may not feel safe challenging or overriding an AI-generated recommendation in the first place. Only 28% of CHROs say HR has a joint roadmap with IT backed by a shared operating cadence, and 73% say they struggle to coordinate consistently across the rest of the C-suite.
That coordination gap has a measurable cost. IBM found that organizations which explicitly define which workflows are human-led, AI-assisted, or fully AI-executed report an 18% reduction in operational risk and a 20% improvement in output quality, compared with organizations that leave the boundary undefined. Yet 46% of organizations do not involve the CHRO at all when AI strategy is being set, which helps explain why so few workflows carry that clarity.
Skills erosion is not a hypothetical for workers who are already worried about it
Among the employees who told IBM they worry about skills erosion, three out of four said AI has already begun eroding at least some of their skills, not that it might someday. Skills erosion also ranks among CHROs’ own top concerns, cited by 46% of respondents. The two groups converge on which skill matters most: 49% of employees and 57% of CHROs both name critical thinking and problem framing as the capability that counts most in an AI-saturated workplace. Where organizations build judgment checkpoints directly into how AI-assisted work gets done, 62% of CHROs report growing employee confidence in AI-enabled decisions. Where judgment is left out of the workflow, 57% of CHROs report confidence declining instead.
“AI is changing not only how work gets done, but where people can contribute the greatest value,” said Nickle LaMoreaux, Senior Vice President and Chief Human Resources Officer at IBM. “As AI takes on more routine and process-driven tasks, uniquely human capabilities become even more important. CHROs have a critical role to play in redesigning the workplace of the future so people can focus on the areas where they can have the greatest impact.”
The productivity gains are creating a second, unrecognized job
IBM’s study also surfaces a pattern that HR leaders elsewhere in the industry have started calling a second shift: work that AI creates rather than removes. Eighty percent of CHROs in the study believe AI adoption creates what they describe as “invisible” work for employees, including validating AI recommendations, fixing its mistakes, supplying missing context, and managing exceptions the system cannot handle. Forty-two percent of employees confirm it from the other side, saying AI has increased their workload or that the added work goes unrecognized by their organization. Only 42% of CHROs say the productivity gains AI generates are being reinvested into innovation or reskilling, meaning the majority of that freed-up capacity is going somewhere else, or nowhere measured at all.
The accountability structure compounds the problem. Where the CHRO at least shares responsibility for deciding which calls stay human-led, 76% of employees say they feel safe questioning or overriding an AI recommendation. Where HR’s role is merely advisory, that figure drops to 43%. The difference is not a matter of technology maturity; it is a matter of who employees believe has their back when they push back on a machine.
What it means for the HR leader
The study lands hardest on HR’s own house. Seventy-two percent of organizations make limited or no use of AI inside the HR function itself, even as CHROs are asked to govern AI everywhere else in the business. CHROs rate their own teams’ capabilities low across the board: 13% call HR’s AI literacy adequate, 16% say the same of HR’s ability to measure AI performance, and 20% rate HR’s change-management capability for AI adoption as sufficient. IBM found that organizations with mature HR AI capability, spanning governance, architecture, and measurement, are nearly twice as likely to report a higher number of positive business KPIs than those without it.
The practical takeaway is that judgment cannot stay an unstated assumption. HR leaders who want the confidence numbers to move in the right direction need to name, in writing, which decisions in a given workflow are human-led, which are AI-assisted, and which are fully automated, and then make sure IT and the rest of the C-suite are working from the same document. Without that clarity, the study suggests employees will keep absorbing blame for AI’s mistakes, keep doing invisible work that never shows up in a productivity dashboard, and keep losing confidence in tools their own leadership is telling them to trust.
HR also cannot credibly ask the rest of the business to build AI governance it has not built for itself. The 72% of organizations where HR barely uses AI in its own function is the same population reporting the lowest confidence in its ability to measure or manage AI adoption elsewhere. Closing that gap, more than any new tool purchase, is what IBM’s data suggests will move the CHRO and workforce numbers back into alignment. HRTech Edition has previously reported that AI governance capacity is lagging deployment speed across HR broadly, and this study quantifies exactly where that lag is doing the most damage: in the judgment calls closest to individual employees. It also echoes concerns raised in recent reporting on AI’s effect on the leadership pipeline, where the same erosion-of-judgment dynamic shows up one level up the org chart.
Source: IBM Newsroom
