A joint ESRI and Department of Finance report published this week estimates that AI could displace around 7% of Irish jobs — roughly 200,000 positions. Unlike previous waves of automation, the risk is concentrated in highly educated, middle-to-high income roles. Here is what the findings mean for leadership teams.
A joint report from the Economic and Social Research Institute (ESRI) and the Department of Finance, published this week, contains a finding that will be uncomfortable for many Irish leadership teams: the workers most exposed to AI-driven displacement are not the lowest-paid. They are the people in the room.
The report — Artificial Intelligence and Income Inequality in Ireland — estimates that around 7% of current Irish jobs could be displaced in the short to medium term. That is approximately 200,000 positions. But the distribution of risk overturns the conventional narrative about automation and employment.
What the report actually found
Previous waves of technological disruption — manufacturing automation, computerisation — primarily displaced routine, lower-income work. AI disruption is different. The ESRI analysis shows that the highest displacement risk sits in roles that involve structured cognitive tasks: information processing, document production, numerical analysis, communication coordination.
The occupational categories with the highest projected displacement:
- General and keyboard clerks: 18% displacement
- Numerical and material recording clerks: 15.8%
- Customer service clerks: 14.6%
- ICT professionals: 13.7%
- Business and administration professionals: 11.4%
- ICT technicians: 10.6%
These are not entry-level roles. They are the backbone of how most Irish professional organisations operate: the people who manage information, coordinate processes, produce documentation, and provide technical expertise. They are also disproportionately represented in financial services, professional services, the public sector, and technology — sectors that between them employ a substantial share of Ireland's workforce.
Why this is different from previous automation
The reason this pattern differs from historical automation is the nature of AI's capability. Previous automation replaced physical or highly repetitive cognitive tasks. AI language models are effective at tasks that require synthesis, drafting, summarisation and structured analysis — exactly the tasks that define many mid-to-senior professional roles.
A junior analyst who spends 40% of their time producing reports, summarising data and drafting presentations is not doing irreplaceable work in an AI-enabled environment. A senior manager who coordinates information flows, produces governance documents and manages structured correspondence is similarly exposed.
This does not mean those roles disappear immediately. It means that the value proposition of those roles — and by extension, the justification for the headcount — changes under AI adoption. Organisations that deploy AI effectively will get the same output from fewer people in those categories. That is a productivity gain. It is also a workforce management challenge that most Irish leadership teams have not yet confronted directly.
What the ESRI report says about inequality
The report's income inequality findings are nuanced. Wage gains from AI adoption are "modest but broadly shared" — AI raises productivity, and some of that flows to workers. But those gains are "not large enough to counterbalance the average fall in income due to job displacement" at the economy-wide level.
The distributional finding that matters most for leadership teams: middle and higher-income households are more exposed than lower-income ones. This is the reverse of historical automation, where technology compresses inequality by displacing low-wage routine work and increasing demand for higher-skilled roles. AI, at least in its current configuration, appears to compress in the opposite direction — reducing demand for structured cognitive work that sits in the middle of the income distribution.
Ireland's tax and welfare system is assessed as relatively well-placed to absorb the shock for lower-income households through automatic stabilisers. The exposure sits higher up the income distribution — in roles and households that are not typically the focus of social protection policy.
What this means for leadership teams right now
The ESRI report is an economy-wide study. Its implications for individual organisations are more specific and more immediate.
Your team's current composition reflects a pre-AI world. The balance between human effort and AI-assisted output in roles involving document production, data analysis, client communication and information management is shifting rapidly. The question is not whether to act on this but at what pace and with what governance.
The roles most exposed are often the roles with most institutional knowledge. The ICT professional who has deep system knowledge, the senior administrator who knows how every process actually works — these are not purely fungible. Displacement statistics are averages, not certainties. The organisations that navigate this well will distinguish between roles where AI is a genuine substitute and roles where the human element is irreplaceable even if the task mix shifts.
Workforce planning needs to catch up. Most Irish organisations are still running workforce planning frameworks built for incremental headcount growth or reduction. AI-driven role change is neither — it is qualitative change in what a role requires, alongside potential quantitative change in how many of those roles are needed. That requires a different analytical approach.
The risk of inaction is not zero. Organisations that do not deploy AI in roles where it creates genuine leverage will face a productivity disadvantage relative to competitors that do. The ESRI findings should not be read as a reason to avoid AI adoption — they should be read as evidence that adoption decisions require serious governance, not ad hoc tool deployment.
The leadership team's own exposure
There is a particular conversation that most leadership teams are not having, which is about their own roles. The ESRI data suggests that ICT and business administration professionals — the category that includes many senior operational roles — face 11-14% displacement risk at an economy-wide level. The implication for senior leaders who spend a significant proportion of their time on coordination, communication and information management is not comfortable.
An AI-enabled leadership team does not need the same overhead structure to produce the same governance output. That is a board-level workforce planning question that is easier to ask now, before external pressure forces it.
If your organisation wants to understand where its AI exposure sits and how to manage the transition deliberately rather than reactively, a structured diagnostic is the starting point. At Acuity AI Advisory, we work with leadership teams to map actual working patterns, identify genuine AI leverage, and build a governance approach that is proportionate to the risk. The work starts with evidence, not assumptions.