ESRI researchers address the Oireachtas Joint Committee on Artificial Intelligence

On 10 November, ESRI researchers Dr Karina Doorley, Dr Dora Tuda and Dr Tara McIndoe-Calder addressed the Oireachtas Joint Committee on Artificial Intelligence.

Drawing on ESRI research, they outlined recent research evidence on AI and its impact on work in Ireland.

Read their Opening Statement

Introduction

Let me begin by thanking the Chair for the invitation to the ESRI to appear before the Committee. I am Karina Doorley, and I am joined by my colleagues Dora Tuda and Tara McIndoe-Calder. We are grateful for the opportunity to appear before the Committee today to discuss AI and work.

AI is beginning to reshape the kinds of work people do, how incomes are earned, and the balance between labour and capital in modern economies. For Ireland, a country with a highly educated workforce and a strong technology sector, AI adoption presents a unique challenge and opportunity. I will now briefly discuss the results of three pieces of ESRI research that relate to AI and work.

Recent ESRI research evidence on AI and work on Ireland

In a joint report with the Department of Finance, my co-authors and I explore what AI adoption among firms in Ireland could mean for Irish households over the short- to medium-term.

Using evidence from international research, we simulated a range of plausible scenarios for AI adoption and its effect on employment, wages and capital income. In our central scenario – drawn from a US study – around 7 per cent of current jobs could be displaced in the short–medium run. Those remaining in employment experience an average wage increase of 2.6% due to increased productivity and the return to capital increases by 0.4 pp. It is important to note that this is a scenario and not a prediction. It is likely to be an upper-bound scenario as recent evidence from Europe points to smaller potential employment and wage effects.

In all scenarios, those most likely to experience job displacement are found in higher income households, where the share of workers transitioning into unemployment is substantially larger than in lower income families. For those who remain in work, we estimate broadly shared wage gains. Increases in returns to capital because of AI adoption benefit households at the very top of the income distribution, where the vast majority of Ireland’s capital income is concentrated. The combined effect on household disposable income depends on the relative balance between job loss and wage gain. In scenarios where household income losses occur, Ireland’s tax and welfare system absorbs most of the income loss for lower income households, and roughly half of the loss for households at the top of the income distribution. 

Income inequality, measured by the Gini index, rises moderately in every scenario we examine due to the polarising effect of job losses and wage and capital income increases on the income distribution. 

In joint work with the Department of Enterprise, Tourism and Employment, my colleague Tara McIndoe-Calder and her co-authors have carried out two pieces of forthcoming research exploring the impacts of AI on the economy. 

The first paper explores the potential impact of AI on the labour market in Ireland in the context of substantial structural changes to occupations over the last two decades. The research finds that employment in the occupations most exposed to new digital technologies has been falling, in both relative and absolute terms since the early 2010s, substantially before AI as we now know it.  The rate of switching between occupations with different levels of exposure and complementarity to digital technologies is also rising over time. Additionally, there are relatively high rates of skills changes within occupations with both high exposure to and high complementarity with digital technologies. This implies that even occupations already requiring digital skills are seeing rapid changes in skills requirements. 

The second paper explores the recent adoption of AI technologies by micro, small and medium enterprises (SMEs) in Ireland. There are four main takeaways from the research. First, there is evidence of a digital divide between the AI-adoption rates of large and smaller firms. Second, SMEs have substantially increased their adoption of AI and their intentions to invest in these digital technologies over a short period. However, over 70% of SMEs are either yet to adopt AI and/or are not willing to invest further in AI. Third, regulatory uncertainty and security concerns have become prominent obstacles to investment, especially by firms that have already adopted AI tools. Finally, firms that use AI more intensively are substantially more likely to have formalised its use within the organisation and are also more likely to report positive experiences with the technology.

Conclusion

These findings underline the importance of forward-looking policy. AI has the potential to increase productivity and living standards, but only if the workforce is equipped with the skills needed to use it effectively or to transition into less exposed roles. Ireland’s high levels of educational attainment offer a strong foundation, but targeted educational support will be essential, particularly for older workers or those with lower formal qualifications.

In relation to domestic firms, whilst the evidence points to continued and broad-based diffusion of AI technologies among Irish SMEs, the results also highlight the importance of addressing barriers to adoption if firms are to translate initial adoption into sustained investment and more effective organisational use. Knowledge-related barriers remain strongly associated with both AI adoption and investment intentions. This suggests that awareness-building and information diffusion continue to play a critical role in supporting adoption. Similarly, the importance of skills shortages, regulatory concerns and implementation challenges points to the need for policy measures that support firms beyond the initial adoption decision.

Our analysis also highlights the need to consider the future resilience of Ireland’s tax base. If AI accelerates the shift from labour income to capital income, the current heavy reliance on labour taxation may become increasingly difficult to sustain. Broadening the tax base – as advised by the Commission on Taxation and Welfare - and strengthening taxation of wealth and capital may become necessary to ensure the long-term stability of public services and welfare supports.