A CFO uses an AI assistant to draft a board update before the Monday morning meeting ends. Down the hall, employees performing similar knowledge work remain stuck with old processes because access, training, managerial support, and incentives never arrived. That divide exposes the real AI adoption gap: companies can appear technologically advanced while most daily work remains unchanged.
Research from the National Bureau of Economic Research makes that problem difficult for HR leaders to ignore. Drawing on responses from almost 6,000 senior executives in the United States, United Kingdom, Germany, and Australia, the researchers found that 69% of firms actively use AI. Yet executives average only 1.5 hours of AI use per week, and nine in ten report no employment or productivity impact during the previous three years.
Access is spreading much faster than meaningful adoption.
Executive Expectations Are Outrunning Workplace Reality
Executives expect AI to increase productivity by 1.4% over the next three years while reducing employment by approximately 0.7%. Employees surveyed separately expect employment to rise by 0.5%. That disconnect creates an immediate AI workforce planning challenge.
HR must connect executive forecasts with decisions about hiring, skills, performance expectations, job architecture, and internal mobility. When leaders privately anticipate workforce contraction while employees hear only optimistic messages about empowerment, trust deteriorates as soon as requisitions disappear or vacant roles remain unfilled.
The productivity evidence should also temper corporate enthusiasm. Most firms have yet to see measurable results, suggesting that software licenses alone will not deliver the anticipated AI productivity gains. Organizations need to identify which workflows should change, what authority AI systems should receive, where human judgment remains essential, and how managers will evaluate AI-assisted work.
Employment reductions may arrive quietly. Rather than announcing mass layoffs, companies may cancel requisitions, delay backfills, shrink graduate programs, and leave entry-level roles vacant. Research from the Stanford Digital Economy Lab adds urgency to the potential AI hiring slowdown. Workers ages 22 to 25 in highly AI-exposed occupations experienced a 16% relative employment decline after researchers controlled for firm-level shocks.
That pattern raises a long-term talent question: Where will companies develop experienced professionals if they automate the work through which beginners traditionally learn?
HR Must Redesign Jobs Without Destroying Career Paths
The answer begins with task-level analysis rather than sweeping predictions about entire occupations. The International Labour Organization estimates that one in four workers holds an occupation with some exposure to generative AI, with clerical roles facing the greatest exposure. However, exposure does not mean every job can or should disappear.
Effective AI job redesign separates routine production from judgment, relationship management, creativity, exception handling, and accountability. Some tasks can be automated. Others can be accelerated. Still others should remain human-led.
This analysis gives HR a defensible basis for determining where to reduce hiring, redeploy capacity, create new responsibilities, or preserve developmental work. An entry-level assignment that appears inefficient may still teach employees how the business operates. Removing it without creating an alternative learning pathway can weaken succession pipelines years later.
HR should therefore review job families, promotion criteria, workforce plans, and learning programs together. Employees need opportunities to practice AI-enabled work inside real workflows, receive coaching from trained managers, and demonstrate new capabilities through projects that affect advancement.
The World Economic Forum reports that 63% of surveyed employers see skills gaps as a major barrier to transformation. A credible AI skills strategy must define proficiency by role, provide protected practice time, and connect learning to mobility rather than treating training as an optional library of videos.
Trust and Measurement Will Determine Adoption
The expectations gap between executives and employees makes workforce participation essential. OECD workplace research found that training and worker consultation are associated with better employee outcomes following AI implementation. Building employee AI trust requires direct answers about how jobs will change, which decisions will involve AI, what information systems will collect, and how employees can challenge errors.
Employees do not need promises that every job will remain untouched. They need credible commitments about transparency, fair evaluation, learning opportunities, and access to emerging roles.
HR also needs stronger measurements. Adoption rates reveal who opened a tool, not whether work improved. Effective workforce AI metrics should track cycle time, quality, rework, employee confidence, skill growth, customer outcomes, internal mobility, and differences across demographic groups.
These safeguards matter because AI increasingly affects recruitment, evaluation, monitoring, and other employment decisions. The Equal Employment Opportunity Commission warns that automated systems can violate federal anti-discrimination laws when used in workplace decisions. Responsible AI adoption therefore, requires vendor scrutiny, accessibility testing, human review, escalation procedures, and regular analysis for adverse outcomes.
Closing the distance between executive expectations and employee experience will require an AI transformation strategy that treats job design, capability building, communication, and governance as business infrastructure.
The organizations that gain the most from AI will not be those with the most licenses. They will be the ones whose employees understand how work is changing, possess the skills to succeed, and trust the workforce decisions being made around them.
Gleb Tsipursky, PhD, serves as the CEO of the future-of-work work consultancy Disaster Avoidance Experts and wrote “The Psychology of AI Adoption at Work: From Resistance to Results” (Georgetown University Press, 2026) and “Returning to the Office and Leading Hybrid and Remote Teams” (Intentional Insights, 2021).