A new IBM report has surfaced a concern HR leaders have flagged: artificial intelligence adoption may be eroding the very human capabilities that chief human resources officers say matter most to organizational performance.

The IBM study, covered by HR Dive, reframes a familiar workplace worry. The conversation about AI has long centered on job displacement. IBM's argument runs in a different direction: as AI handles routine cognitive work, the humans reviewing or approving that output stop practicing the underlying skills themselves. Capability declines because the work is no longer done by hand.

CHROs surveyed in the IBM report identified the affected capabilities as central to leadership and organizational learning. That framing puts the issue inside HR's traditional remit, not the CIO's.

What does the report change for HR?

The finding forces a re-examination of workforce plans that treat AI as a productivity lever alone. If generative tools quietly deskill a workforce, the projected return on AI investment is partly offset by a deterioration in the human capability the technology was meant to augment.

HR teams will need to pair AI rollout plans with deliberate practice structures that maintain proficiency in areas the technology now handles. That requires investment at a moment when most learning budgets are tilting toward AI literacy itself.

Which capabilities are most exposed?

IBM points to a class of skills that resist easy automation: synthesis across large information sets, negotiation, ethical reasoning under ambiguity, and the ability to coach and develop others. These are the same competencies HR has spent the past decade building into leadership pipelines.

The risk is not that AI performs these tasks poorly. It is that AI performs them well enough that humans stop practicing. A manager who lets a model draft every performance review loses the facility to write one unaided. A recruiter who relies on algorithmic ranking stops forming independent judgments about candidate fit.

What should employers track?

The practical question for HR leaders is measurement. Most organizations have dashboards for AI usage and productivity. Few track skill atrophy in the workforce itself. Three indicators are worth watching now:

• Time spent on at-risk tasks without AI assistance
• Rolling twelve-month assessments of skill quality by peers and managers
• Share of high-stakes decisions still routed to human judgment rather than model output

CHROs who treat AI deployment as a standalone technology decision rather than a workforce-capability decision will inherit both outcomes: the projected productivity gains, and the loss of the muscle needed to govern those gains.

Where this goes next

The IBM report is likely to surface in board-level AI governance discussions that have so far focused on data privacy, model bias, and hallucination risk. Skill erosion belongs on the same agenda. HR leaders who can quantify the trade-off will shape how their organizations deploy AI over the next planning cycle. Those who cannot will find the conversation set by vendors and engineers instead.