Who Is Exposed When AI Arrives? Map the Risk Before It Maps You
AI exposure is uneven across job types. SMEs that identify it early can protect people and productivity at the same time.
AI is creating jobs and putting others at risk in the same period. LinkedIn research reported by The Guardian shows AI job postings have doubled since 2023 and pay on average more than twice as much as roles without AI work (Pardes, 2026). The same research finds that women are more likely to hold roles with high exposure to AI disruption, such as customer service (Pardes, 2026).
AI exposure is uneven across job types.
The pattern matters beyond gender. Roles built on routine, high-volume tasks are the most exposed, and they exist in almost every SME: customer service, administration, data entry and basic reporting. When the people in those roles are also concentrated in one demographic group, an AI rollout can widen existing pay and progression gaps.
The article also shows that holding an AI job is not enough. Women in AI roles were reported to cluster in lower-paid work such as data annotation. Across all AI occupations, men’s median pay was $45,000 higher, partly because of the jobs each group tends to hold (Pardes, 2026).
Implications for SMEs
Map exposure by role. List roles by how much of the work is routine and rule-based. This is a talent risk and a transformation risk in one exercise.
Check who sits where. Review the composition of exposed roles and of any new AI-related roles. Uneven patterns are an early warning.
Build pathways, not just tools. Offer paid time and structured training so exposed staff can move into higher-value work.
Document decisions. Record how roles were assessed and why. This supports fairness, and it supports your duty of care when work changes.
AI adoption can lift productivity. It lifts it further when the people closest to the work are brought along rather than replaced by default.
Limits. The figures come from a LinkedIn report as summarised by a news outlet, and the data reflect a largely US technology context. SMEs in other sectors and countries should test these patterns against their own workforce data.
References
Pardes, A. (2026, 8 October). Women hold just a fraction of new AI jobs but are overrepresented in roles with high risk of AI disruption. The Guardian.
