IBM Report: AI Use Risks Eroding Skills CHROs Deem Essential

What happened

IBM has published a report cautioning that growing use of artificial intelligence in the workplace risks eroding human capabilities that chief human resources officers (CHROs) consider essential, according to HR Dive. The report adds to a wider body of commentary questioning what heavier reliance on AI tools may be doing to employees' cognitive skills and judgement over time.

HR Dive notes this is not the first warning of its kind: concerns about AI's effect on employee thinking and skill development have surfaced repeatedly as organisations expand adoption of generative AI and automation across day-to-day work.

Why it matters

For leaders driving AI adoption, the finding reframes the conversation from pure productivity gain to workforce sustainability. If staff increasingly offload reasoning, problem-solving or communication tasks to AI tools, organisations may inadvertently hollow out the very human skills — judgement, critical thinking, interpersonal nuance — that CHROs flag as core to leadership pipelines, service quality and resilience.

This has direct implications for experience design and workforce strategy: AI roll-outs that are not paired with deliberate skill-building risk weakening the human layer that differentiates service in moments AI cannot fully handle, from complex complaints to judgement-heavy decisions.

The Renascence take

The instinct in most transformation programmes is to measure AI success by efficiency metrics alone — tickets resolved, hours saved, cost per contact. This report is a reminder that the real cost of automation can be a slow, invisible decline in human capability that only shows up when it's needed most.

Skill atrophy is a behavioural problem disguised as a technology one: people stop practising what they stop doing. Any organisation deploying AI at scale should treat human-skill retention as a designed outcome, not a hoped-for side effect — building in deliberate "no-AI" reps, coaching loops and escalation scenarios that keep judgement, empathy and critical thinking active. The operators who get this right won't just have faster service; they'll have people who still know what to do when the model gets it wrong.