Human judgment is emerging as a central workforce concern as organisations adopt AI, according to a new IBM study announced on September 21, 2026. The research, conducted by the IBM Institute for Business Value with Oxford Economics, compares what senior workforce leaders say companies need with what employees fear AI could take away.
The study combines responses from 1,500 CHROs and equivalent senior executives across 21 geographies and 23 industries with a separate survey of 8,800 full-time employees in 28 countries. Its findings point to a clear divide: leaders emphasise the need to control AI systems, while employees are more concerned about losing skills through their use.
Executives put oversight ahead of automation
In the executive survey, 71% of CHROs identified the ability to supervise, validate and override AI outputs as the most essential workforce skill in the AI era. Critical thinking followed at 57%, while human judgment was selected by 48% of executives as an important capability.
Employees gave human judgment less weight. Just 29% listed it among the important skills for working in an AI-enabled environment. At the same time, 60% said they were concerned that AI could erode their skills, with critical thinking the skill most often identified as being at risk of decline.
That difference does not show that AI has already caused critical-thinking skills to deteriorate. It does show a tension employers will have to manage: leaders may view oversight as a capability to develop, while employees may experience the same change as a threat to abilities they already rely on.
Clearer workflow roles correlate with better reported results
The research also connects explicit workflow definitions with reported operational improvements. Organisations that distinguish between human-led, AI-assisted and AI-executed workflows reported an 18% reduction in risk and a 20% improvement in quality.
Those figures describe the organisations’ reported outcomes. They do not establish that the labels themselves caused the reduction in risk or the improvement in quality, and the study does not show how the results differ by industry or geography.
The relationship between AI governance and HR remains uneven. Forty-six per cent of organisations do not involve the CHRO when defining their AI strategy, while 72% use little or no AI within their HR function. Companies are therefore identifying human oversight as a priority while many have not yet made the workforce function a central part of their AI planning.
The practical issue is accountability
IBM’s findings give organisations a concrete framework for discussing AI adoption: identify which work remains human-led, where AI assists, and where AI executes, while keeping responsibility for validation and override clear. The survey does not prescribe a universal policy, but it shows why those boundaries matter to both implementation and employee confidence.
For workers and employers, the central question is not simply whether AI is present. It is how much judgment remains visible, accountable and protected within the workflow. The study suggests that critical thinking and human oversight are becoming workforce priorities even as many organisations remain at an early stage of applying AI to HR and broader operations.