We have received apologies from Senators Laura Harmon and Lynne Ruane and Deputy Sinéad Gibney. Substituting for them will be Deputy George Lawlor, Senator Alice-Mary Higgins and Deputy Pádraig Rice, respectively.

I advise members of the constitutional requirement that they must be physically present within the confines of the Leinster House complex in order to participate in public meetings.

We will suspend for a few moments to allow our witnesses to join us.

The evidence of witnesses physically present, or who give evidence from with the parliamentary precincts is protected by absolute privilege in respect of the presentations they make to the committee. However, they are expected not to abuse this privilege and it is my duty as Chair to ensure that this privilege is not abused. If their statements are potentially defamatory in relation to an identifiable person or entity, they will be directed to discontinue their remarks.

It is imperative they comply with any such direction. Members are reminded of the long-standing parliamentary practice to the effect that they should not comment on, criticise or make charges against a person outside of the Houses or an official either by name or in such a way as to make him or her identifiable.

Our meeting this morning is a discussion on AI and work. It is the first in a series of meetings we will have on this very important topic. We have all heard about the potential for AI to create job displacement and losses. I hope we will also hear if AI can benefit workers or create new, meaningful roles in Ireland. We are joined in person by representatives of some of the main trade unions in Ireland and also online by a member of the European Trade Union Institute and Ms Catherine Keogh from Fórsa. From the Irish Congress of Trade Unions, we are joined by Mr. Owen Reidy, general secretary. We are also joined by Dr. Tom McDonnell, co-director of the Nevin Economic Research Institute. From Fórsa, we are joined by Mr. Ryan McKinney, national secretary for health and welfare division and the Fórsa lead on AI policy, and Ms Catherine Keogh, who is online. Joining use from the Financial Services Union are Mr. Brian McDowell, head of communications and public affairs, and Mr. John O'Connell, general secretary. We are also joined by Ms Molly Newell, technology researcher at TASC. Joining us online is Dr. Aude Cefaliello - I hope I pronounced that properly and please correct me if I did not - who is a senior legal researcher on health, safety and working conditions with the European Trade Union Institute. Finally, from the Communications Workers' Union, CWU, we are joined by Mr. Seán McDonagh, general secretary, Mr. Ian McArdle, deputy secretary general, and Ms Fionnuala Ní Bhrógáin, head of organising and campaigning.

We will start off with ICTU. I invite Mr. Reidy to give his opening statement on ICTU's behalf.

I thank the Cathaoirleach and the committee for holding this event today. As the committee probably knows, congress is the representative body for 46 unions and associated bodies on the island of Ireland. We represent about 750,000 workers, North and South, making us the largest civil society organisation on the island. I am joined by Dr. Tom McDonnell from the Nevin Economic Research Institute, NERI.

We believe job quality should be the primary goal when we look at AI because it will clearly change our workplaces, and in potentially very radical ways. To be clear, the trade union movement is not a block to the appropriate development and spread of AI. However, workers need to be central to the debate and the decisions about how the technology is developed and applied. We know the technology has profound implications for productivity growth over the medium to long term and, if managed correctly, it could lead to a boon for living standards. The hype is certainly alive and well but AI advancement is not an organic or neutral process. There is nothing inevitable and nothing that happens innocently. How AI develops and the implications for workers will depend crucially on the decisions of policymakers.

Job quantity is regularly portrayed as the greatest risk from AI, yet all the historical evidence we have from previous job disruptions from technology is that job quality is where the greatest impact is to be found over the long term. A concern is that catastrophising over job losses will undermine attempts to improve job quality. There will always be job losses and job creation. This is inherent to the creative destruction of technological and institutional change, and we have seen it over the decades. However, we can shape this trend in the direction of job quality via our policy decisions, for example, in education policy, labour market and enterprise policy, tax policy, regulatory policy and so on. The policy focus must be on maintaining and improving job quality, both for the jobs that will be retained and the new jobs that will be created.

AI will replace some tasks within many occupations. It is well recognised that routine cognitive roles, such as clerical and legal, are most at risk of disruption, while non-routine manual roles, such as farming and plumbing, are least at risk. In contrast to the three previous industrial revolutions, the disruption will manifest most heavily in white collar sectors and tasks. There is risk of attrition through reduced hires of younger cohorts and non-replacement of retiring staff. How the remaining human component within jobs interacts with or complements AI, and what tasks remain, is the key to understanding the impact on workers.

AI will make certain jobs harder and others easier and will have different impacts for early-stage versus established career workers. We do not think existing studies of potential job losses are particularly helpful. AI adoption should not be presented as a single event but as a process which will last decades. While there is some evidence of short-term hits, there is, as yet, no real econometric basis for most estimates of job losses and at this early stage, estimates are highly speculative.

My colleague, Dr. McDonnell, is happy to discuss in more detail that underlying labour market assumption arising from AI but from a policy perspective, we emphasise the need to focus on technology for exploration, in other words, for new businesses and ideas, rather than technology for exploitation and simple cost savings.

An additional concern are the potential changes in the distribution of returns to labour and capital leading to a rise in inequality, job polarisation, loss of agency and voice for workers, as well as rising income inequality. If these outcomes were to come to pass, we are looking at a breakdown in social cohesion. Ireland is particularly at risk from potentially transformative changes given the high level of knowledge-intensive and professional jobs here. Ireland could be a canary in the AI coal mine if AI policy is mishandled, with particular concerns for entry level and junior employees. On the other hand, we have high levels of STEM graduates and a high concentration of AI talent, which leaves us well placed to potentially benefit.

How should we prepare for the changes? Social cohesion is a precious thing, and it should not be taken for granted. We need workers' voices and their trade union representatives front and central to policymaking. AI will have impacts throughout society and not just for enterprise. If there was ever a reason for this State to award and give workers the right to collective bargaining, its time has come with AI. We should also acknowledge this by transferring AI policy from the Department of enterprise to the Department of the Taoiseach so that a more holistic approach can be adopted as part of a proper social dialogue. As part of that improved dialogue, we need to conduct ongoing strategic foresight exercises and establish sectoral task forces. We need to reform our education and skills policy with greater emphasis on life-long adult learning. This can be paid for through the National Training Fund.

We will need to redesign social insurance taking an approach based on the just transition principle to adequately protect the incomes of workers and employers, with a view to better labour market matching. We will need to rebalance the composition of taxation away from labour and towards capital owners. We will need to regulate malign use, protect privacy and redesign AI biases and discrimination. We will also need to strongly regulate against algorithmic management. Finally, we need to pursue work-life balance policies with flexible hours and strong labour protections to free adults to pursue education alongside employment.

We have failed with technological shifts in the past. We cannot hold back the tide, but we can shape its course. We are happy to take questions.

Go raibh maith agaibh for the invitation to address the committee on AI and work. I am joined online by my colleague, Catherine Keogh. Our union, Fórsa, represents more than 100,000 workers across the civil and public service and the private sector.

I will begin with this committee's own work. Its second interim report recommended that trade union negotiation be a precondition for AI-driven change in public services and that the guidelines for the responsible use of AI in the public service be made mandatory. Fórsa supports both of those recommendations, and our central message today is simple: they can and should be acted on now. Wider deliberative processes, including the citizens' assembly the committee's report proposed, can complement action in the workplace but must not become a reason to defer it.

This summer, in a statement titled "We Must Act Now", more than 2,000 economists and AI researchers - 17 Nobel laureates among them - warned that AI's transformation of work could exceed the industrial revolution's but compressed into a fraction of the time. Bill Gates, who only three years ago called this transition "manageable", now writes that it will be "one of the most turbulent times in human history", and that AI will either be "the greatest equaliser ever invented, or the worst source of injustice".

The first major study of AI at work on this island, which involved 4,300 workers and was published last month by University College Dublin, Queen's University Belfast and the Nevin Economic Research Institute, found that almost a third of workers now use AI on the job. Fewer than one in 20 has seen any increase in earnings from it. A third say their work has intensified. Most who save time simply absorb more work. The gains from this technology are real, but they are not being shared. Almost half of Irish workers now believe AI will replace or damage their role, up 19 points in a single year.

What reassurance do workers get from the State, the largest employer in the country? The Minister for public expenditure has said that the era of "endless recruitment everywhere" is over, and that services can be delivered with fewer staff.

When the HSE published its AI for Care strategy, it consulted a stakeholder group of more than 30 organisations. Patients were consulted through a citizens' jury. On the other hand, representatives of the people who would work alongside these systems every day were not consulted at all.

Let me be clear about Fórsa's position: we do not stand in outright opposition to this technology. People are not afraid of the tools but of the terms. We stand for frameworks that are negotiated before systems are deployed and applied to the systems that are already in use. We stand for no AI adoption premised on cutting jobs. We stand for a named, accountable human in command of every consequential decision. We stand for the sharing of the benefits of AI's productivity gains through time and flexibility, pay and reinvestment in services.

Where AI takes over the routine parts of a role, what remains is the harder, more human work, which involves judgment, care and relationships. That must be recognised in grading and pay and not used as grounds to reclassify the role downward.

None of this is utopian. European law already provides for collective agreements as the vehicle for governing workers' data. Nor is it a minority view. Eight in ten Irish workers told Eurobarometer that they want workers and their representatives involved in the design and adoption of new technologies. The workers' voice is not a brake on progress. It is what makes progress shared and sustainable.

At our conference in Killarney in May, An Taoiseach told us that we need to go further than the current model of social dialogue. He was right. Ireland has done this before. When the ground shifted beneath this country in 1987, the response was not a seminar; it was a negotiated framework that endured for two decades. Faced with this challenge, the question is whether we have the foresight to build arrangements equal to the pace and scale of what is coming. Our position paper, which we have shared with the committee, sets out our proposals in further detail.

Working people will judge this moment in history not by strategy documents and summit speeches, but by the impact on their payslips, workloads, schedules and services. The test is whether the people who do the work are treated as authors of what comes next or as an audience for it. We look forward to the committee's questions.

The Financial Services Union is the leading trade union representing staff in the financial services, fintech and tech sectors. We support employees of retail banks, global financial services organisations, technology and fintech organisations, and also workers in the games and animation sectors. The FSU thanks the committee for giving us the opportunity to present to it.

Artificial intelligence is expected to reshape patterns of work in Ireland's financial and technology sectors. While these groundbreaking technologies offer opportunities to improve productivity and services, there are significant concerns around job security, changing skills requirements, workplace surveillance and inequality.

FSU's position is clear: AI must work for people and not the other way around. We believe there is a clear correlation between potential efficiencies and productivity gains for employers from responsible AI adoption and greater reward, including flexibility in ways of working, for our members. With an estimated 97% of financial service and insurance jobs exposed to AI, ensuring workers share in the benefits of this transformation is an urgent priority.

Last week, the Irish Government hosted the international AI summit as part of the EU Presidency. The keynote speaker at this flagship event was Mr. Brad Smith, president of Microsoft. In his address, Mr. Smith specifically identified trade unions as key allies in the adoption and integration of AI. Hand in hand with change of the scale predicted from AI must come guardrails for workers. The most important aspect would be workers having a real say in the AI adoption process through collective bargaining.

A 2025 survey by TASC and the Financial Services Union highlighted this tension of workers asking what is the upside for them. While respondents did report some perceived productivity gains, 88% of respondents expected AI to contribute to job displacement and only 28% felt adequately prepared to use AI tools at work. These findings underline that the impact depends not only on the technology itself, but on how it is implemented and governed.

Ireland's policy approach thus far has prioritised digital adoption, competitiveness and skills development, with less emphasis on socio-economic equality and workers' participation.

There must be a workers' dividend from AI, more flexibility, smarter ways of working and a just transition in the real sense of the words. Greater worker consultation is critical to ensuring that productivity gains translate into better outcomes for employees rather than simply greater profits for employers.

Some 37% of Irish workers’ core skills are expected to change in the period from 2025 to 2030, according to the World Economic Forum. Over that same period, 41% of skills in financial services and capital markets and 47% in insurance and pensions are expected to change. This is higher than the global cross-industry average of 39%. AI and big data, technological literacy and cybersecurity are expected to be among the fastest growing technical skills, alongside creative thinking, resilience and curiosity. There is a significant gap between these changing requirements and workers’ current preparedness. Only 28% of FSU members report feeling prepared to use more AI tools in their daily work. Only 15% of FSU members report that their employer has implemented initiatives specifically designed to develop skills for integrating AI. More than 40% say their organisation has offered no such training and has no plans to do so, while 31% are unsure whether training is available at all.

As of September 2026, the FSU-Bank Of Ireland agreement is the only collective bargaining agreement in Ireland directly addressing AI. This agreement establishes principles covering transparency, human oversight, data protection, social dialogue, reskilling and job protection, and provides that significant decisions affecting employees should not be made solely through automated systems. The agreement was also informed by a survey of union members incorporating workers’ perspectives into AI governance. The FSU-Bank of Ireland agreement demonstrates how collective bargaining can establish practical protections around transparency, human oversight, training and job security. Given the benefits that workers' involvement in AI adoption-related decisions can bring to organisational roll-outs, employers should be equally interested in fostering meaningful opportunities for dialogue. The FSU calls for stronger worker involvement in AI adoption decisions, expanded access to training and Government policies that centre around workers’ rights and well-being. Ireland’s financial services and tech sectors can remain competitive while protecting workers but achieving both requires giving employees and their representatives a meaningful role in shaping the future of work in Ireland.

I thank the committee for inviting me. It is an honour to contribute to this important discussion on artificial intelligence and work. I am senior legal researcher at the European Trade Union Institute in the working conditions and health and safety unit. My work focuses primarily on occupational safety and health law, including the implications of new technologies and artificial intelligence at work. The remarks I will make also draw on the work of colleagues at the institute.

My central message is simple: AI at work is not only a question of innovation, productivity or competitiveness. It is first and foremost a question of work organisation, workers’ health and safety and industrial democracy. It is important to remember that AI at work is much broader than algorithmic management or the platform economy. AI can automate physical tasks, for example, through collaborative robots; cognitive tasks, particularly through generative AI; managerial tasks; and, in some cases, even preventive tasks. AI-based systems can provide early warnings of safety and health risks, analyse the effectiveness of preventive measures and suggest improvements. Automation may also be partial rather than complete. This diversity matters because the risks and opportunities of AI depend on what it is used for, how it is deployed and who controls it.

Today, I would like to focus in particular on the risks that AI-based worker management systems create for psychosocial health and safety. This is where the evidence is most concerning and where the regulatory response remains insufficient. Algorithmic management and digital monitoring can create or exacerbate psychosocial risks.

They can intensify work, increase workload and work pace, reduce workers’ autonomy and make working time more unpredictable. They are also strongly associated with overconnectivity, work extending beyond normal hours, frequent contact outside working time, work performed in free time and constant availability. These dynamics contribute to stress, exhaustion, anxiety, job insecurity and work-life conflict.

Performance monitoring can also increase workers’ fear of losing their jobs. It can contribute to the dehumanisation and datafication of workers and to the use of private or sensitive data. In some cases, workers may feel forced to choose between following automated instructions, meeting productivity targets and protecting their own safety. The pressure to work faster can lead to more repetitive movements, awkward postures and reduced attention to ergonomics.

These risks are not only psychological. If an AI-based worker management system malfunctions, relies on inaccurate data or directs workers towards a hazardous situation, the consequences can include serious physical injury and, in extreme cases, death. This is particularly relevant in sectors such as manufacturing and warehouse work, where interactions between workers and vehicles or machines can be dangerous.

AI does not have to produce these outcomes. Worker representation and social dialogue can play a decisive role in shaping how these technologies are introduced and used. Employees have first-hand knowledge of how work is actually performed, including the practical challenges, needs and risks that may not be apparent at the management or design level. Involving workers and drawing on their perspectives and insights in decision-making can help ensure that technologies are better adapted to real working conditions and production processes, increasing the likelihood of successful implementation and acceptance. Research shows that greater worker involvement is associated with technologies that are better adapted to production processes, higher productivity and more moderate negative effects on employment. Worker participation can also help identify, prevent and mitigate psychosocial risks arising from AI-based worker management.

This is why information, consultation and collective bargaining should not be seen as obstacles to innovation. They are conditions for a safer, fairer and more effective deployment of AI at work. Workers and their representatives need access to understandable information about the AI systems used in their workplace, including how they function, what data they collect and how they affect work organisation, performance evaluation and health and safety.

The current European legal framework does not adequately meet this challenge. The AI Act does not sufficiently address the full range of workplace dynamics, occupational health and safety risks, and the collective dimension of AI deployment. The platform work directive addresses only a specific category of work - platform work - and does not provide a general framework for AI and algorithmic management across the labour market. At the same time, psychosocial risks remain insufficiently addressed at the EU level. This is why clear legal provisions are needed on employers’ duties when deploying AI at work. These should include occupational risk assessment; transparency on the functioning of AI systems and the collection and use of data; effective information and consultation of workers and their representatives; and safeguards to ensure that AI is used to support workers rather than intensify control and subordination.

I wish to make a specific call to Ireland, which holds the Presidency of the European Council from July to December 2026. The European Commission’s quality jobs Act is expected by the end of this year and its consultation has covered algorithmic management and AI at work, as well as health and safety risks, including psychosocial risks. Ireland is, therefore, uniquely placed to ensure these issues are addressed together and not diluted in the forthcoming negotiations.

If we want to benefit from the positive potential of AI, we must prevent its negative effects, especially for workers’ health and safety, and give a real voice to the workers who will live with these technologies every day. I thank members and welcome their questions.

I thank the Chair, Deputies and Senators for the opportunity to address the committee. The Communications Workers Union represents workers in the postal and courier sector, including An Post; the telecommunications sector, including Eir; and, increasingly, digital and technology workers through our own digital and tech worker alliance branch.

Our growing membership in tech gives us a direct view of AI and its more immediate impacts. These workers build, train, maintain, moderate and use digital systems.

They understand both the technology and the claims made about it as well as the way it is used to justify thousands of redundancies in the sector.

Through our robust collective bargaining structures, where we have them, we can manage transformative change where we represent workers in a manner that deals with their concerns fairly. In doing so, we can ensure that the change is more meaningfully embedded. However, AI differs in its speed, its reach into professional and creative work and its capacity not merely to automate tasks but to monitor, measure and manage people. So far, workers have been almost entirely absent from the conversation. The public debate has been dominated by competitiveness, investment and productivity. The Government summit last week gave barely five minutes to the lone voice from the Irish Congress of Trade Unions - my colleague, Ms Fionnuala Ní Bhrógáin - and it is our view that we cannot have a serious discussion about the future of work without the people who actually do the work.

The future of AI is not predetermined. It is not set in stone. Employers, unions, workers, communities and, most importantly, government have choices to make that will define the deployment and development of AI. The discussion on AI must also ask who keeps their job and we should not pretend that there will not be job losses. What happens to those who remain, who controls the technology and, critically, how do we share the dividends? For our members, this is not hypothetical. We have seen hundreds of tech sector job losses or jobs placed at risk in restructuring processes where AI and automation have been cited. AI is not the sole cause. Outsourcing, cost-cutting and corporate decisions also matter. These are choices made by employers, not inevitable decisions made by machines.

Nor is the threat limited to redundancy. AI is already capable of allocating shifts and tasks, measuring output, tracking workers and influencing recruitment, performance management and dismissal. Unchecked by the balance of worker input and collective bargaining, AI will be used to intensify workloads, undermine professional judgment and erode dignity. A worker should never have to accept that AI or a large language model is responsible for a decision affecting his or her livelihood. There must be human accountability and a human in charge.

There is also evidence that the gains are not being shared. Research presented at our Workers’ AI Counter Summit last weekend, based on the Working in Ireland Survey: An All-Ireland Study 2025, found that 29.3% of workers surveyed in the Republic were using generative AI at work. Among users, 33.5% reported a faster or more intense pace of work and just 4.7% reported higher earnings arising from its use. These are self-reported findings but they raise an urgent question. If AI saves time, why are workers not seeing more of the dividend? Workers and communities need transparency on the true cost of AI in terms of data centres, climate impact and the potential erosion of democracy as well the cost itself to the business. Adoption of AI, especially by small and medium businesses now, may lead to excessive cost burdens in the future as AI companies seek to actualise their profit. Investors in AI will demand a return on their investment, and businesses and workers will end up paying the price.

That takes me to our central concern - collective bargaining. Ireland still has no general statutory requirement for employers to recognise trade unions and bargain with them. The Government’s Ireland's Action Plan to Promote Collective Bargaining 2026–2030 is a welcome acknowledgement of the problem, but it reiterates a clear commitment to the voluntarist model. A worker can have the right to join a union, while the employer retains the right to refuse to negotiate with that union. In an age of rapid AI deployment, that is a serious barrier to fair and effective change.

In reality, our voluntarist model is a veto model. We have seen this time and again in collective redundancy situations. If there is no union recognised by the employer, the staff elect their own representatives who then enter a 30-day consultation period with a view to reaching agreement on the issues arising from the redundancy. Early this year, we saw what these measures mean in reality: worker proposals for ex gratia payments rejected; pleas for minimum payments for staff with less than two years’ service refused; profitable employers exercising a veto; and hundreds of workers let go from work with no redundancy pay left, in their own words, “jobless, penniless and homeless”. Why do employers behave like this? Because they can. Why do employers refuse to collectively bargain with staff? Because they can. Our country is, shamefully, in a class of its own as the only western EU state that does not provide a legal right to collective bargaining and as a country, a democracy and an economy, we are weaker for it. Informing workers about a system is not the same as allowing them to negotiate whether it should be introduced, how it should operate and what happens to their jobs and conditions.

We are asking the Oireachtas to act on three priorities. The first is to legislate for effective trade union access, recognition and collective bargaining rights, including negotiation of significant workplace AI changes. Second, we ask that enforceable safeguards be established; advance notice, joint impact assessments, transparency over monitoring and automated decisions, independent scrutiny and meaningful human review. Third, we require a just transition, with paid training, redeployment, protection against avoidable redundancies and a fair share of productivity gains through pay, working time and improved conditions.

Let me be clear that the CWU is not opposed to technological progress. Our union has navigated successive transformations in the sectors we represent. We have been dealing with AI in its earliest form, which was machine learning, for many years in telecommunications. When workers collectively bargain, change can happen in a positive and productive manner to the mutual benefit of all stakeholders. The question before us is not whether AI will shape the future of work. It is whether workers will have the power to shape that future too. I thank the committee.

Before inviting questions from members, I thank the witnesses for being here at what is an extremely busy time for them all. We appreciate their taking time out of a hectic schedule and with everything that is happening at the moment.

I will call members according to the speaking rota, with each having five minutes for questions and, I reiterate, for answers. I ask, please, that they try to keep to the five minutes. I am conscious that we need to finalise our interim report on AI, truth and democracy after this session. I hope there will be time at the end of the first round of questions for one or two further questions from each member. We will then have to move to discussing the finalisation of the interim report.

Senator Harmon is first on the rota. Deputy Lawlor is substituting for her. He has five minutes.

I thank the witnesses for their presentations. We often hear it said that something is barrelling down the tracks towards us. In the case of AI, it has well and truly driven over us and, in many respects, we are playing catch-up. I come from a printing background and am old enough to remember the days of Wapping and the introduction of new technologies. Workers' rights certainly were not front and centre in that area. It was deskilling and down-skilling that were front and centre of a lot that went on at the time. I see that happening again in the whole area of AI. We can be guaranteed that the benefits will not be destined wholly to go to workers. There is no doubt that AI has huge benefits, particularly in the area of health. I acknowledge that.

My first question is for Mr. O'Connell. I am interested to hear about the agreement with Bank of Ireland. As he stated, it is the first such agreement. Will he flesh out how it came about, how it is working and the union's interaction with the company?

Second, I am hearing from talks on the subject that we really need a stakeholders' forum on AI. How might that look and who should organise it?

Regarding the Bank of Ireland agreement, we have existing change agreements with the bank. Indeed, we have existing technology change agreements with it. The difference with AI, as referred to by Mr. McArdle, is that both the pace and scale of the change are such that they require a revised approach. Other things have developed over the years, particularly in terms of surveillance and issues like that. People want safeguards. They understand that we are in a changed environment but they want safeguards around that. We engaged with our members and conveyed their views to the bank.

The origins of the agreement come from Europe. I sit on the social dialogue committee for the banking sector in Europe, at which a set of principles on AI were agreed. We brought that set of principles to Bank of Ireland, noting that they had been agreed by the bank's representative bodies in Europe and stating that we believed they provided the basis for us both to reach agreement on how change and AI should interact in Bank of Ireland.

It addresses some things that are normal in change agreements, but issues like human oversight, transparency and surveillance are relatively new in the workplace in the form that they come to us with AI. We are also looking at the human-in-control principle because that is really important and, again, was not something we envisaged a number of years ago when we were bringing normal change agreements through the organisation. That is the Bank of Ireland. It is a template in terms of change and how change can come about. As everybody on the trade union side has said here today, we are not opposed to the change; it is about how it comes about, what influence we can exert to shape that change and our entitlement to do so. That is part of it. In relation to the stakeholder forum, Mr. McDowell may wish to come in.

I thank the Deputy for the question. We were before the finance committee two or three years ago, looking for a banking forum which would bring stakeholders together throughout the ecosystem in banking to see how we could manage the change that was coming in the financial system. A hugely successful report was done by the Department of Finance with 34 different recommendations. Legislation for access to cash was one of the items agreed through that banking forum. It brought together consumer groups, NGOs, trade unions, employers, the Government and every relevant stakeholder. It is our firm view that we should do something similar here because it is going to have such a serious effect on the workplace.

We met with the Minister of State, Deputy Smyth, who is responsible for AI, and asked her if she would consider it. It is our view that her Department should be the one that calls those stakeholders together because at the end of the day, and I am repeating what others have said, we are not against the integration of AI into the workplace, but it is about how that is managed. It has to be done collectively and collaboratively. For the workers, it has to be done transparently. Getting relevant stakeholders in a room is a good way to start. If it is done properly, it will be good for workers, businesses and consumers. Let us get them all in a room together to discuss it.

I thank the witnesses for their opening speeches. There were a lot of similarities in those speeches.

Mr. McArdle summed it up quite well when he spoke about the importance of job quality. That resonated with a lot of what was said in the other contributions. We have been engaging with a lot of witnesses for quite a while now. Worker consultation has come up fairly repeatedly, as well as human decision-making. Today I am hearing about shared productivity gains and opposition to workplace surveillance. Collective bargaining is a message that has come through clearly from all the witnesses. A lot of people have mentioned the potential for job losses but Mr. McArdle put it quite well when he said that, at this early stage, estimates are highly speculative. It is very important to say that. Deputy Lawlor said it is a train flying down the tracks at the moment, but it has been doing that for a long time. When I graduated 22 years ago, technological upskilling and cybersecurity were going to be the things we all had to focus on, and that was an opportunity for us. Some 22 years later, there is still a huge opportunity. Mr. McArdle said this process will last decades. It is something that is constantly evolving.

Mr. McKinney acknowledged the previous report this committee put together. In a lot of the areas around worker consultation and human decision-making, we have heard those points loud and clear. They are to the fore of everything we are doing at the moment. It was only this morning that the ESRI published a report on healthcare that states Ireland's growing and ageing population will challenge the financial sustainability of the Irish health and social care system.

Here, we have a technology that is going to take a huge amount of pressure off that. I come from County Roscommon. We have people travelling from Donegal down to Roscommon to get healthcare services. These are three- or four-hour journeys. We would not be doing our jobs if we did not ensure we are in a position to try to take advantage of the opportunity to evaluate information as it is coming in and better inform doctors and patients and remove that need. That is an exciting part of it.

A few of the witnesses mentioned upskilling, the just transition and areas like that. That is very important. Of all negotiations, that is the key thing that we need to be focusing on. In education at the moment, we are talking about inclusive education and staff who have their qualifications for 20 years who are crying out to be given time to upskill and train. The onus is going to be on the unions to ensure that the National Training Fund is taken advantage of and they draw down whatever they can. It is very important that is regionally delivered, not just focused on the urban areas. There is a bigger opportunity.

The training is a big thing. The FSU is the first one to have included this in a contract. The whole idea of addressing AI within an employment contract is such a dynamic area. Mr. O'Connell mentioned change about five times in his opening few words. What his union has done is the key on it. I was going to ask Mr. O'Connell to expand a little bit more on it but Deputy Lawlor got in there before me. I refer to transparency, human oversight, data protection and social dialogue. On the reskilling, if Mr. O'Connell would not mind giving me a quick overview-----

Mr. O'Connell is only going to have 30 seconds.

I know that. He has 43 seconds.

Reskilling and upskilling are vital. As Mr. Reidy said, there is the catastrophic view and then there is the complementary view. Nobody knows where it is going to land. If it is in the scale of change that we think it is, the issue of upskilling is vital. Investment needs to be made. There are models that we discussed that can provide people with the right skills to retain the roles. Some people will want to move on and make a career decision and so forth, but the vast majority of people will want to be retained and will want to be upskilled. That is an investment in protected time, so people have that time to gain the appropriate skills. People need to be forward planning and skilling for the future.

The next speaker is Deputy Murphy.

I thank all the witnesses for being here. I wish to start with the CWU. Significant credit should go to the CWU for the lead it has taken in this area. The workers summit on AI was excellent and an important contribution. Ms Ní Bhrógáin was expected to single-handedly provide balance in a five-minute contribution at the AI summit against 36 big tech and big business CEOs and management. I have a question on the AI summit. When did the invitation to the AI summit come to the CWU or ICTU?

To the best of my knowledge, I was asked about a week and a half, possibly, before the event. I am happy to double check that for the Deputy.

That would have been after all the other speakers were announced.

Yes.

It was after we had an engagement here with the Minister of State, Deputy Smyth, on 23 September, when she was questioned about that there were no trade union speakers.

Again, I am happy to double check the dates. To the best of my acknowledge, the full programme had been announced before we were invited.

I thank Ms Ní Bhrógáin. As someone who was at the event, what was Ms Ní Bhrógáin's perception of whether this was a balanced debate where workers' viewpoints, civil society, climate and so on were appropriately represented in the discussion?

My impression - I am referencing back to my colleague's opening statement - was that the tone of the entire event was centred around investment, competitiveness and innovation. That carried through with all the speakers. In essence, it was about Ireland Inc. and making ourselves as palatable as possible for further future foreign direct investment. There was very little, if any, consideration given to the impacts of other voices.

Certainly, ours was the only trade union voice on the day and there was no other consideration in regard to civil society.

Thank you. I will now turn to Dr. Cefaliello. Will she expand a little on the important point she made about AI at work being “much broader than algorithmic management or the platform economy”. There is the automation of physical tasks but also a role in terms of management and, for workers who work with their hands and their bodies, like warehouse workers, the potential danger of AI mistakes and so on and even the potential for serious injury or death. Can Dr. Cefaliello expand on the sort of danger and what that looks like in reality? Have we seen any cases of that so far?

On the automation of physical tasks there is a brilliant study from EU-OSHA report that shows the deployment of collaborative robots, also known as cobots. These are robots with intelligence and capable of autonomous going around the workplace. At the beginning it was actually scary for the workers because there was no security perimeter and the robots were running around with screwdrivers, for example. We asked the workers. First there was a deployment and a first assessment. While the deployment of the technology happened, we asked workers about how they lived with it and how we could improve it, especially the stress, anxiety and different psychosocial risks, autonomy, control and all these things. That led to better deployment. There are concrete cases in different countries. I can send the committee the EU-OSHA report.

I spoke about increasing AI but the key thing is management. In warehouses there is often a sort of gamification so there is work intensity. As a result of that work intensity and because workers are not in control that can lead to increased stress and cardiovascular problems such as heart attacks, which is something that happened a couple of years ago in one of the warehouses in Poland. A worker died from a heart attack because of the intense working conditions and the use of algorithmic management. That is why it is important that when we discuss AI we need to ask what we are discussing: is this collaborative robots? Is it algorithmic management? What are the risks, how do we consult and how do we evaluate?

That was also a striking feature of the workers’ summit – the role of increased intensification both for physical labour and mental labour.

I have a final question for Dr. Cefaliello. She pointed out that AI in work is not adequately addressed by the EU AI Act or the platform work directive and there is need for further regulation. Obviously the main direction of travel in Europe is the other way, in terms of deregulation through the digital omnibus and so on. What role is the Irish Government paying in any of these things currently?

Dr. Cefaliello is not able to see the amount of time remaining. If she does not mind, can she answer in 30 seconds, please?

As I said, there is the Presidency of the Council and there is discussion about the quality jobs Act where there might be legally binding provision regarding algorithmic management, AI and psychosocial risk that could benefit workers. Ireland could act on this and also adopt its own legislation to be a model for other countries and show a good example.

This is the most witnesses we have had and everyone is very welcome. I will start with the Irish Congress of Trade Unions. In his opening statement, Mr. Reidy stated, “workers need to be central to the debate and the decisions about how this technology is developed and applied”. What really came through from all the witnesses was that the most important thing is collective bargaining. Without collective bargaining, where will this end up? As things stand, does ICTU believe employees and their representatives are being given a meaningful role in shaping the future of work in Ireland and developing policies?

The Deputy is 100% right. Without collective bargaining we are at a loss here - this will not be done right and it will lead to social cohesion breaking down, not today or tomorrow but at the next crash. We will see job displacement done in a way that is not without the engagement of people and workers. As regards how I would characterise it at the moment, the AI summit was a classic example. I can tell Deputy Murphy we got a call on 28 September for a trade union nominee.

That is how late in the day it was. It was very much a token gesture after the event. Ireland's attitude there is to see how we can make a quick bob. It needs to be much more holistic and it needs to be much more fundamental. As Mr. McArdle said, the employers do what they do because they can. They need to be compelled to engage with workers when workers freely choose to join a union. If we do not get this right, I fear for societal cohesion and I do not think we should take it for granted.

There will be issues as happens in all labour markets. Jobs come and jobs go. There are bad jobs that we would be happy to see go. We want to see high-skilled, high-paid unionised jobs that are effective and efficient where workers can codesign, be innovative and be productive. However, at the moment they are being shut out. We have a couple of positive examples which some of the unions have highlighted. That is because the employer in that instance engages. The vast majority of employers choose not to and the State needs to compel them.

The adequate minimum wages directive will not do it unfortunately. The action plan will not do it. The LEEF AI forum that the Taoiseach has announced is welcome but it will not do it. We need AI moved out of enterprise. It should not just be seen from the narrow view of how we can make a quick buck here and needs to be a much more societal thing. We need the Government - a government - to once and for all grasp the nettle of collective bargaining. We will not see a flight of capital out of the country. Where will they go? Will they go to the Netherlands, Belgium, France, Spain or Germany where they do collective bargaining? There is nothing to fear here. Without collective bargaining we are in real trouble.

I have a question for Dr. Cefaliello. What would she like to see included in the European legal framework? What is it missing regarding work organisation, performance evolution and health and safety?

We need to once again broaden the understanding that we have for AI, not only management but also the other forms so it will be future-proofed. There needs to be an obligation regarding transparency about how the technology is used and what data are collected. There is a huge part about being sure that workers understand and are consulted about the deployment and how data are used. Then there is a huge part about evaluation of the physical and psychological risks in the deployment and the use. That should be done at the moment of the deployment but also in a circular manner. The thing with AI is that it evolves depending on the data it collects on the ground. We do not completely know the way it will evolve. It will already be a huge step forward to have this recognised, to force dialogue on this and to have legal provision on that.

I have a question for the CWU. Everyone has spoken of a just transition. I was on the agriculture committee and we talked about just transition. For example, a farmer going from milking to tillage would need to have three or four years for that transition to happen. How will this transition happen when AI is developing at so rapid a pace?

As Mr. O'Connell and Mr. Reidy mentioned earlier, reskilling and retraining have to be part of that conversation. The most effective way to deliver on that is through collective bargaining in the workplace. Ideally, we would have a national framework that sets out the rules of engagement. AI is happening and, as Deputy Lawlor said, it is steamrolling through workplaces. We have seen it lead to the redundancies across the tech sector. We have not seen that in the wider industry yet. There has to be a just transition that deals with training and redeployment.

We have gone over time and I ask Mr. McArdle to wrap it up.

I thank the witnesses. Regarding just transition for current workers, do the unions have any comment on what they are seeing with recruitment in the first instance at entry level? Do they have any concerns on that?

Some of the evidence here is international. There does seem to be some softening of entry level jobs in professional services, consultancy firms and those types of things. There have been significant increases in hiring in recent years in areas such as tech, mostly associated with Covid, and a ramping up, so many of the job losses we have seen in recent years in many ways relate to a correction from the hiring during that period. In certain areas, such as computer programming, there have been reduced hires in recent years and we see a decline in employment in those areas, but the reality is that the data is still much too early. We are only, let us say, three years into a process that will play out over at least three decades.

The argument that is made by the employer is that we continue to recruit but, essentially, workers are going up the value chain.

Yes. One of the issues with failing to continue to hire is that eventually those lower level workers will become your mid-level and higher level workers, so it is actually a false economy in the longer term for these firms where, essentially, they hollow out their skill base that is coming through in the years ahead. I think there is a genuine debate within these sectors as to the best way forward and there is a little buyer's remorse in terms of some of the non-hiring that has been going on.

We would say that it is likely that the automation channel will dominate in the short to medium term. There will be displacement effects. Obviously, productivity effects will have positive and negative implications as to what that means for jobs but, ultimately, what we have seen with these technologies is that the innovation channel tends to dominate. The way we might see it is that in the shorter term there might be a downward correction in terms of labour demand but over the medium to longer term - of course, in the long run we are all dead - that should wash out. That was what we saw over the three previous industrial revolutions. The difference here is that we have much more evidence and much more understanding about how these processes play out so there is greater scope to catch it early in a way we did not with the previous industrial revolutions, and indeed globalisation and all those major changes that led to huge job shocks.

The FSU wants to come in on that as well.

I thank Deputy O'Rourke for the question. When we speak about hiring, it is also important to think about the use of automated services within this and the encoded biases that are within these systems. It is not just the early evidence we see in the softening of the labour market for young workers. These systems have been consistently found to disadvantage female workers, workers of colour and workers coming from non-conventional backgrounds. Speaking of the pipeline, the increased use of these technologies could result in a less diverse workforce going forward. It is important to think not just about how it is softening the number of people we are hiring, but also about how it is affecting the types of workers we are hiring and how that could undermine efforts at culture and diversity within workplaces.

If I may, I will ask the CWU a follow-on question about collective bargaining. The argument in Ireland on the employer side is that we have this voluntary code and we do good work in that regard. I ask the CWU about the experience of that and, maybe, without putting words into the witnesses' mouth, why that is not sufficient.

It is not sufficient because we have seen time and again workers seeking representation collectively through their union. To speak specifically about the organising work we have done in the tech sector, at the start of this year we had the first strike in the tech sector in Ireland. Our union undertook that on the basis that there were 700 redundancies being threatened in a contracting company supporting one of our major social media platforms. Those people were undergoing a collective redundancy and, as I mentioned in my statement, any of their proposals to mitigate the redundancies or to improve their terms and conditions were ultimately rejected by the business. Where there is no collective voice, workers are disenfranchised, workers are weaker and the outcome of any change agenda being implemented or imposed on a workplace will not be as effective. Collective bargaining is good for workers, good for the economy and good - essential, in fact - in any change agenda involving AI to be effective in the workplace.

I thank the speakers. I am interested in the point about quality of work. There is one point on quality and one point on quantity and I would like to hear the thoughts of the witnesses on them. On the quantity piece, we are told there will be swings and roundabouts and there is employment coming in other areas. Last week, however, we heard that a large amount of the employment in areas such as data centres is construction employment and we already have construction employment. In fact, it is displacing workers from areas where we have identified shortfalls in the availability of construction workers. I am sure ICTU will cover this. It is not new employment; it is a displacement of existing workers. That is balanced against real job losses in areas where we have seen the loss of employment. How important is it that we have that reality and accuracy on the quantity piece?

On the quality side, which Mr. Reidy was right to focus on, just weeks ago, Jeff Bezos was talking about the wonderful new future with a shorter working week. People pointed out that people are dropping down on the floors of his factories because they are denied toilet breaks. Will the witnesses talk a little more about what is crucial in the framing of quality, especially when we see this algorithmic management that Dr. Cefaliello spoke about? Will the witnesses comment on the danger of algorithmic management in terms of the deterioration in the quality of employment?

Many companies are actively moving away from diversity, equality and inclusion, DEI, policies and being required to do so in the United States. Some of the same companies operate here or are lending the technologies used here in terms of management, and they are very actively working against equality measures, as Ms Newell pointed out. How does that sit with the hard-won gains on the gender pay gap and those aspects of the quality of work we have been working on?

My last question is a little wider. It relates to other workers, namely, those producing precious minerals overseas. I attended a meeting earlier today with indigenous people from Guatemala who are against mines. We also know this happens in West Papua and the Philippines. There is an extremely aggressive new era of mineral mining. Has there been any links between unions here in Ireland, unions on one end of the tech AI chain, and workers further down the chain who are experiencing extremely brutal impacts of the new race for AI?

On quality and quantity, in this country we measure too much on quantity and not enough on quality. An example is the crazy decision to give the VAT cut to hospitality. What does that do? It means there is public policy to sponsor poorly paid and bad-quality jobs. They are important but they are not the future or the answer.

On quality, there are a couple of very simple elements. Worker voice is one - independent voice - and the opportunity to co-design health and safety, the capacity to have demands upon you that are reasonable and decent terms and conditions of employment. It is not all about pay; job quality is about many other factors as well. I will pass over to the other witnesses.

There will be a lot of substitution of jobs but complementarity will be important too. The same technology can make certain jobs harder and others easier. The effect will not be evenly spread. Even within the same occupation, replacement of some tasks will have different effects on different workers, whether it is early stage versus established career. There will be winners and losers, even within the same firms. The reality is that with previous technological changes, there were declines in living standards which have lasted for decades, for example, in the first industrial revolution and other technological advancements. The only way we can prevent this is by ensuring workers are part of the process. Left to themselves, employers do not care about workers. It will benefit them.

I would also like to point to the income inequality that could arise from this-----

Ms Newell has five seconds to answer.

-----where high-paid jobs go, low paid jobs are created, there is greater income inequality, a falling labour share of income and, obviously, an impact on social cohesion, political power and so forth.

I will speak briefly to some of the gendered impacts of AI within the workplace. I referenced the issues with bias in hiring systems. It is also important to think about the types of jobs that are more exposed to AI. Those are more likely to be in clerical roles and administrative work in which women tend to be over-represented, meaning that in order for them to advance in their careers they will need be more active in terms of upskilling. This will create an additional burden on women when they are more likely to already face burdens at home relative to housekeeping and their male partners. It is also important to recognise that research has shown that women have fewer advanced digital skills and feel less prepared to use AI. This was confirmed in our survey of financial services workers as well. This is the same even when one corrects for labour and education. It is important, therefore, to think about upskilling and how we can support women without burnout.

I thank all of our witnesses and welcome the proactive approach the trade unions are taking. In particular, I compliment the FSU. I have referenced the FSU's approach in a sector that has been exposed to a lot of the changes that we are seeing. The approach is that AI is here, it has been here and it is rapidly transforming everything we are doing. In fact, one of our difficulties on the training side, which I will come to, involves keeping pace with the change that is happening.

Mr. O'Connell referred to the World Economic Forum report on the future of work. One of the interesting statistics in that report shows that globally, we are looking at about 92 million jobs being displaced between now and 203 but potentially 170 million new roles being created. Globally, we are going to see a huge churn, so it is really about how we prepare for that.

My argument has been that it is going to be increasingly difficult to protect specific job roles because of the nature of the change and what we have to look at is how we can sustain human livelihoods. An over-emphasis on protecting titles would be difficult because the transformation is going to be huge. We are moving into a situation where information is cheaper and analysis and operational tasks will increasingly be carried out by AI but judgment then becomes a more valuable commodity. We have to look at things like contextual ethics and relationship intelligence - those kinds of skills that are going to be crucial. The witnesses will be aware that, with the National Training Fund and in other areas, the Government has decided to place a large emphasis on upskilling and reskilling. Where should that be targeted in the AI space?

Ms Newell has already indicated the types of roles that could get a direct first hit, namely, administrative roles. If we are talking about the future of work, a couple of components come into it. It is about the quality of the work that people do but it is also a once-in-a-lifetime opportunity to transform people's working lives. I was on a call yesterday with a number of people in our sector. They spoke about the challenges of balancing work and home and the fact that they adopted all these skills during Covid to work online and now that we are back in or partially back in the office, they have to manage both. They have to manage their work life in the office and their work life online. The people who are in the direct line are identifiable but we are all going to be affected. As Mr. Reidy said earlier, we can adopt two approaches to this. One is a really positive approach where we are supported and there is real engagement in terms of collective bargaining at the workplace level on what the impacts are.

I get it in terms of those who are obviously directly targeted. I am concerned that we are not moving quickly enough. Are there any specific skills?

I will speak not specifically on skills but on concerns about how that education and reskilling happens.

It needs to be consistent with people being in their jobs. We are talking about micro-credentials, which could happen in the evening or at weekends. We are not talking about four-year university courses here. That is too late for somebody in their 40s or 50s, or even three-month or six-month courses. It needs to be a redesign of how the higher education sector is providing those services to the business sector and workers. We really need to rethink how we do this.

As for the specific skills, they are going to evolve quickly over time too. Obviously, a lot of transferable skills will be very relevant here. The Deputy talked about knowledge and judgment. In terms of who is coming through the system, such as teenagers and those in their 20s, again greater emphasis on STEM will continue to be very important.

I have one final question in the context of public sector pay talks. Part of what we looked at here is that AI technology should be seen as a tool and not a goal in itself. It can be used and we have seen plenty of examples of that. Public servants have used AI quite effectively to be able to better deliver public services. I certainly believe a key part of the public sector pay talks has to be the responsible use of AI. In certain areas, for example, public procurement, I certainly favour the use of agentic AI. I know Revenue, for instance, and other organisations are looking at that. As part of the talks, is there an acceptance on the part of the trade unions that AI makes a contribution when it is used safely and responsibly in the interests of the better delivery of public service, which is important to all our citizens?

I ask Mr. Reidy to be brief.

I will be brief because those talks are ongoing and we are not involved. What I will say is that from the work I was involved in at the weekend, when I met both sides, it was very clear that unions were saying AI is happening to us in the public service without us. That is a bad recipe. It is being done to people and not with them. If that continues, people will kick back. There is a rebalancing that needs to take place there.

I thank the witnesses for their time today. I echo the comments of the Chair in expressing appreciation for their engagement with the committee today given all that is going on in their sector. I am hearing the contributions today from a different perspective from some of my other colleagues. Having been an employer, having been self-employed and having run a small business or SME, as the majority of employers across the country do, I understand how important it is to stay competitive and how critical it is that our SME business owners and employers work with their teams to stay relevant and keep pace so they stay viable. I am sure the unions and workers' representatives also understand that.

I will push back a little bit and challenge Mr. Reidy's characterisation of the Government's AI conference and approach thus far as being out to make a quick buck or a quick bob. The Government is absolutely focused on the State staying competitive given that AI is happening on a global scale. It is not happening just within the public sector or within the sectors the witnesses represent. It is happening within the private sector and on a global basis. As the witnesses all know, we are in a very connected globalised economy. Notwithstanding that we want to ensure job quality and ensure fairness and support in a just transition and adoption, it is vital for Ireland, from an economic perspective, that adoption takes place. Upskilling has to take priority, and workers and employers across private and public have to be supported to do so.

Dr. McDonnell said upskilling must happen in a meaningful way. He will be pleased to hear that my colleague, the Minister for Further and Higher Education, Research, Innovation and Science, Deputy James Lawless, more than doubled the budget this year to Skillnet Ireland, a wonderful organisation that helps devise and create customised programmes for employers, in each job and sectoral context, to be rolled out to their teams within their working time. They are often highly funded by the Department. The employer may pay or contribute a small amount to the development of a customised programme. It is potentially only 20% of the cost and places are often fully funded. That budget increased from €25 million last year to €57 million this year. This is a significant contribution and it is a recognition by the Minister that there is a huge surge in the need to keep pace.

I will open it for comments from everybody but I will come to Mr. Reidy first. I was struck by some of the language he used when he said that not supporting workers in this transition - no one is suggesting that we would not support workers in this transition - is a threat to societal cohesion. Having been in the private sector and seen the employer perspective, would Mr. Reidy not recognise those absolutely real pressures to stay competitive in an increasingly difficult and more expensive cost of doing business environment? Would he not recognise that if we create a vast difference between how we treat public sector workers and how private sectors workers are treated and the reality they live in where it is either adapt to AI or die, your business will be viable-----

I ask the Senator to be conscious of time.

Yes, I am nearly there. In its own sense, if we do not treat the workers across both sectors equally, is that not also a threat to societal cohesion?

I spent 20 years in SIPTU representing workers in the private sector where I and all of my colleagues spent all of our days trying to make the firms where our workers were organised, many of them the most cutting edge, high-end FDI firms in this country, to be productive, innovative and successful so our members could share in the success. The reason I characterise it in a slightly flippant way of Ireland trying to make a quick buck is that is our experience and we have been shut out of the conversation. On 28 September, I got a call from the Department of the Taoiseach's office asking if it could have a worker to talk on a panel for five minutes. That is the experience. We are shut out.

Workers who want to raise their voice independently through trade unions are shut out. My concern is that if the Government is interested in productivity, efficiency and innovation, it should unleash the productivity by allowing those workers who want an independent voice to have one. Competitiveness is not just all about cost. The voice does not cost anything. There is an inference that you are adding a layer of bureaucracy if you are discussing, debating and negotiating with someone. I say the opposite: you are getting an efficient and innovative outcome. That is our experience - we are being shut out. If and when the next crash happens and we continue on this course, I worry about social cohesion. What we did last Wednesday was not what is happening in France. We did it in an organised way. By the way, what we did last Wednesday was not what was done by employers in this country in April who flagrantly broke the law. These were farmers and contractors. We do not do that. If we did that, we would be treated differently.

I do not disagree with Mr. Reidy.

My experience is we are shut out, hence the comment it is about Ireland making a quick buck. We need to do it differently and Europe needs to have its own AI. We should not be dependent on America and China. We are up for the debate but we need to be included. The only way we can be included is if the State redesigns the legal and economic framework where unions can be part of the system rather than outside the door. That is the issue.

I am substituting for Deputy Gibney, who is abroad and could not be here. One of the previous speakers talked about their excitement around AI. I do not share it. I have a nervousness, particularly when it comes to the environmental impact, the impact on democracy, data security, workers' rights and social cohesion. There is a lot to be concerned about. That is why this committee and the work is really important. Engaging with unions is really important. If we are going to see the increase in productivity and associated increase in profits, that must be shared with the workers and the people doing the work. I particularly think we need to be pushing reduced working time, such as a four-day work week, as a key dividend for workers if there is to be that increased productivity, as well as increased wages.

There was a discussion earlier about the possible removal of some administrative roles by AI. Is there evidence that workers are now being faced with more complex and draining tasks due to AI while getting the same pay and benefits?

The key point here is the decision on the future travel of AI and what guardrails we put on it will ultimately decide how many jobs there are. In the example of the HSE, which already has an AI for care policy, it talks about a human in the loop. We are saying we need a human in command. A human in the loop removes all responsibility but also removes that job because what is their role if they are only in the loop? If we have decision-makers and people who are ultimately accountable and responsible, those are ultimately human jobs.

Democracy and people in Leinster House need to want that because who do we hold to account if the humans are only in the loop?

It is easy to make generalisations around administrative tasks, for instance. As a trade union, we hold huge amounts of data that we have to manage responsibly all the time. It takes a huge effort to manage that data. To implement AI and use it well in public services requires significant data management, so there are roles in that.

People are taking on more complex tasks and need to be rewarded for that.

They are taking on more complex tasks because there is more complex software to interact with. There are also tasks, however, that will not disappear, such as therapy roles and teachers, for example, that we really need. That is the nitty-gritty that we get into here, but we have to decide first of all that we need human beings in certain roles. Those are the guardrails that we need to put on.

Senator Higgins made a point earlier about the impact that AI-enhanced surveillance and monitoring tools are having on employees in terms of quality of work, workplace safety and well-being. What are Dr. Cefaliello's thoughts on how those surveillance and monitoring tools are impacting on workers?

Before I talk about surveillance, I want to go back to the quality and complexification of the task. To give the committee some numbers from the EU-OSHA polls, when we asked workers, 48% said that digitalisation affects the speed and pace of their work, 19% said that it reduced the potential to use their own knowledge and skills and almost 16% said that it reduced their ability to make decisions about the methods they use or the way they work. This is an important point. With automatisation, we are going to get rid of the simple task and significantly intensify heavy load tasks. So, now, from the beginning to the end, workers are going to do heavy cognitive work when it is used, for example, for translation and services. This is where it is getting complex.

With regard to surveillance, the problem is that workers always feel under scrutiny. There is also the question of rating the tasks. We are going to get rid of the fully automatised ranking and the giving of instructions, but still, when we remove the autonomy of workers, remove the control and treat them a bit like robots, this is where we will have a lot of psychosocial risks. This is also where we need to be sure that the way it is deployed is safe for the workers and that there remains a sense in the way they perform work. I do not know if this answers the questions, but I just wanted to say this.

I really appreciate it. It is really important that those safeguards and guardrails are in place and that the impact on the individual workers is recognised and they are rewarded for it. As noted earlier, we need that right to collective bargaining. It is crucial to us getting there. It has to be a core part of the way we move forward with this in the engagements between workers and employers because there are real risks and concerns that need to be taken seriously.

I thank the witnesses for coming. Hearing from workers' representatives is probably one of the most important sessions we have held as an AI committee. The biggest concern for every constituent we represent is, "What is AI going to do for my job?" It is the thing that every young person in college right now is concerned about. When school students visit me in Leinster House, it does not matter what school they are in, when I tell them I am on the AI committee and ask them what they think about AI, half of the room asks, "Am I going to have job in the future?" and the other half says, "It is kind of exciting and there are great opportunities."

I acknowledge Fórsa, which had a number of us in for a round-table meeting on AI. I found it really instructive and useful. I hope that, going forward, this committee will perhaps be a good one for the union movement to engage with on the role of AI.

One thing we have learned over the last number of months in relation to the public sector strikes is that there clearly has not been enough engagement going on from the political system at large. The less rancour there is in that, the better it is for everyone and for all of our constituents. Mr. Reidy's criticism that he got a last-minute phone call to participate at the AI summit is fair. I do not know how it happened that IBEC was invited to speak at an AI summit and the ICTU or a relevant union representative was not. We should be big enough to just say that it was a mistake and should not have happened.

We should learn from these things, going forward. There should be an AI summit every year. That would be valuable in promoting the opportunities we have here in Ireland through the tech sector that is located here and the 220,000 people who work in the sector and through the Irish start-ups that are building businesses. We could link it with the learnings from abroad in respect of how AI is being introduced in schools in Estonia, Lithuania and other countries, and to learn how labour movements have responded to the introduction of AI in the public sector in other jurisdictions. There are opportunities. We have an opportunity to build bridges.

At the Fórsa event, I got the overwhelming impression that the union movement is not anti-artificial intelligence. When articles are written about some of the comments that have been made today, it might be presented that the union movement is anti-artificial intelligence. Perhaps the witnesses would talk about their own organisations. If we can go around the table, within the witnesses' organisations, is artificial intelligence being used in some form or other? They are all big organisations. What do the witnesses see as the benefits that have flowed from artificial intelligence?

Believe it or not, we are a tiny organisation. We do great work but have only 21 employees. The unions are the large entities.

Of course we use Claude and other AI models in communications, PR and reports. It is a part of life and the research of the organisation. We are not anti-AI. We are against something being done to us. We want it done in collaboration with us. That is the issue and that is what sometimes gets lost in translation. I welcome Senator Ryan's intervention, which exposed that point. We are not anti-AI, but let us do it together in the interests of both sides, rather than just one side, of the labour market.

The Deputy referenced Fórsa. We are all innovative organisations that are constantly striving to improve how we communicate with our members, which is key. Fórsa is using AI in that role. We work with change every day. All of us are involved in managing change and ensuring that our members, the workers, have input into that change. When co-design happens, we see better services. I can think of examples. I am, in the main, responsible for health workers. I can think of examples where services have been improved because teams sat down to figure out how to manage a geographical area or particular priority cases. That happens day in and day out. The same approach will be taken to the roll-out of AI if we are given that role and we need to be given that role.

In Denmark, there is a tripartite arrangement. The unions, Government and employers work together and negotiate. It has been an early adopter of AI. We can be no different but we need to have a voice.

The CWU is also a small organisation. As general secretary, I must recognise the fact that it is foremost a trade union but it is also a business. Equally, it is an employer. We carry all of those responsibilities for our 20 or so workers. Of course we have ongoing consultation with staff. We have training. We provide whatever information, training and skills are needed to assist the business and to assist people doing work. Equally, we must ensure that in the light of concerns, people do not feel threatened by any monitoring of their work situation.

In contrast, if we look at the experience of the people we represent, we have a proven track record in the postal and telecommunications industry in the tech area. The workers who came to us looking for us to assist them had been locked out by their employer and not granted their representation rights. They are foremost and at the front of the early stages of AI and its impact on workers. The response from their employers has been to dismiss and deny their entitlements. The Government has been remiss. When we brought the issue to the attention of the Government, the first response from the Government, six or seven months ago, was that it was monitoring the situation. It did nothing to assist those people.

I remember a time, and I have been a trade union representative for a while,-----

I ask Mr. McDonagh to conclude. We are over time.

------when if employers announced redundancies of up to 1,000, there would be a lot of political activity to try to ensure that those jobs were saved or alternatives were found, rather than silence.

I had some questions but I am going to pivot slightly and stay with Deputy Geoghegan's question. It is interesting to hear it said that of course we use AI and that it is part of life. I would like to probe further into how AI came into the witnesses' organisations. I am assuming it is probably best in class in a way. Fórsa talked about training and said it is part of life, and the CWU is using it as well. Did job descriptions change? At the AI conference, Cisco described how on one Wednesday each month, four hours are set aside specifically for training. How did AI come into the witnesses' organisations? In my constituency office, it started to arrive and we are having conversations about it. We know we need more training. We are nearly a little bit afraid to use it although I know we need more. How did that happen in the organisations present? Did any job descriptions have to change? Was there formal training or did people have to upskill?

We built a new membership database. We now have over 100,000 members and managing that data is key. While we initially thought it was simply a membership database to hold information, very quickly a need was identified to manage that, for example geotagging people, managing their whereabouts, their workplaces and their locations at all times, if we want to communicate with them. We started in a different place, but the journey took us to having to change job descriptions, looking for specific skills. The number of people employed in Fórsa now who are engaging with data and to enable AI is much bigger than we ever thought it would be. That is why I am very confident that in public services, there will be many jobs for people to do. Similarly, it is to enable the AI but we have to have the human element in there. We have the human element and we have very clear guidelines in the policies we have brought in for the use of AI and what data we can put in there. Obviously, we have to work with the commission to ensure that we are constantly managing the data appropriately as well.

Mr. Reidy from ICTU talked about using Claude and so on.

We are a very small organisation with 21 staff, of whom seven are in Northern Ireland and 14 in the Republic of Ireland. It is very small and people are doing a lot of different things. It came in quite organically initially. Different people had different skill sets. There is an ongoing assessment of people's needs so we can see who needs further training and assistance. There is a comprehensive staff manual which is very flexible, as it needs to reflect the changing nature of the times. There would be references in some of the job descriptions, not all of them. Colleagues in communications, PR and research, perhaps, are engaging in AI a lot more than the back-office support staff. Of course, there is finance as well and AI is involved in that too. We are a very small organisation where there is a lot of flexibility. Going back to an earlier point, we are an SME. As Mr. McDonagh says, we are an employer as well and we want to make sure we are an employer of choice and one with the best practice where we engage with our staff and their unions when it is required.

Given that ICTU is a small organisation of 21 staff, how does training look? Is it done on an individual basis? Do people take time out of their day, are they paid or what does it look like?

It is all within their working time. We fund that. It is a matter of doing a skills analysis and finding out who needs what. It is ongoing and usually on a one-, two- or three-to-one basis or in small groups. There are different levels of competency with AI across the organisation, as well as different levels of interest and exposure. I think we are grappling with it in the way we are grappling with all change, and trying to manage it as best we can as a very small operation with 21 employees, two thirds of whom are in the South and one third in the North.

The Bank of Ireland agreement was mentioned earlier.

Is there any mention of reverse mentoring anywhere in that agreement? I am conscious of new graduates coming in or digital natives versus experienced people who have been there a long time. I wonder if that featured anywhere within-----

That question was for the FSU. I ask Mr. O'Connell to be brief.

No, it does not specifically feature but we have done reverse mentoring in organisations, particularly around technology, where young people coming in can train older staff in the skills around technology and take the fear out of it. Equally, older staff can mentor younger staff on policies or progression in the organisation and so forth. We have used it but it is not explicitly called out in that agreement.

Gabhaim buíochas leis an Teachta. Now, I have five minutes. Ar dtús báire, on a personal level, I thank a lot of the witnesses who have engaged with me in the past on different policies I have been writing, particularly on workforce protection, in which I have a huge interest. I hope collective bargaining is part of the talks that are under way. It is a common theme. I would be surprised and shocked if the witnesses did not bring up collective bargaining, to be honest. We need to reimagine the social dialogue or social partnership - whatever we want to call it in the 21st century - and how the State, unions, employers and politicians engage with one another. We need to completely rethink it. If we look at the likes of the Labour Employer Economic Forum, LEEF, for example, does that mean committees are part of it, not just civil servants, and that there is actually more engagement there? I hope that is part of the discussions that are under way.

On Mr. Reidy's nine points about preparing for change, I agree with nearly all of them, apart from one point with regard to preparing for change and the "quick buck". I think we are all trying to get to grips with this and the State does not want to fall behind. I do not think there is any intent on the part of the Government or the State to try to make a quick buck. It is just trying to understand how to do it, but Mr. Reidy and I will agree to disagree on that. I know he was speaking from personal experience but it was the point he was making.

Another point Mr. Reidy made was that AI should fall within the Taoiseach's Department. I raised this with the Taoiseach two weeks ago and his counter-argument was that a lot of people want different things, or perhaps everything, to fall under his Department. Last week, the Taoiseach established a Cabinet subcommittee on artificial intelligence, which he will chair. I hope that will help drive, from the Taoiseach's Department, areas around artificial intelligence.

One of the areas Mr. Reidy raised in his points around preparing for change was the need to rebalance the composition of taxation away from labour and towards capital owners. There will always be some element of taxation on labour. What I am thinking in this regard, and I am not sure it is the same with Mr. Reidy, is taxation on data. Taxing the amount of data that is being stored in data centres or other places might be the future of how our taxation model works. Was that what Mr. Reidy was getting at or what are his thoughts on that?

It is funny that the Chair would say that. I heard a former Taoiseach on a well-known podcast making that very point - that we should be looking at that. I do not disagree. What we were looking at more was the report of the Commission on Taxation and Welfare, which provides for about €16.5 billion in that regard. We think there should be a Sláintecare approach to that and whoever is in Government should be picking at it piece by piece to broaden the tax base. We should be taxing wealth. We have a very progressive tax system, and rightly so. The more you earn, the more you pay, but we do not tax wealth in the same way. There are people with what are considered decent incomes but they are actually quite poor in this economy. It is about taxing wealth as opposed to taxing labour.

Does Mr. Reidy think the local property tax is a way of addressing that? I refer to a tax on a property where people with a bigger house, by and large, have more wealth.

That is one way of addressing it but there are many ways of addressing it. All the talk about inheritance tax is missing the point. Everyone talks about €400,000 or the-----

I have to try to set a good example to everybody else on time. I thank the CWU for the work it has done, particularly around Covalen. We saw with Twitter a number of years ago the lack of collective bargaining. A lot of contracts - I am not specifying any company here - prohibit workers in the tech sector from joining unions. It is a really archaic and dangerous practice and we need to call it out. Everybody should have the ability to join a union.

In regard to the Covalen case, I know there are a lot of roles around trust and safety, which is a very difficult job. It is mentally draining and difficult. From what the CWU has learned, particularly from Covalen, what protections should we put in place to stop that type of incident happening again?

We had Covalen, but then Meta actually laid off staff as well. In total, we had over 1,000 lay-offs within a five- or six-week period. From what the witnesses have learned, what protections should we be looking at for workers?

At the risk of repeating what the Cathaoirleach has heard a few times now, from all the speakers on the trade union side, is-----

Collective bargaining.

-----collective bargaining. It is as simple and straightforward as that. The workers we had organised in Covalen voted democratically to take industrial action. They were on strike for four days. We had engaged with the employer several times and had been rebuffed on every occasion. The Taoiseach actually said on the floor of the Dáil that Covalen should engage with the Workplace Relations Commission and talk to the union. It did not answer any correspondence from the Workplace Relations Commission and it did not engage with us or the workers' representatives.

As I have outlined, what we have is a veto system, and it is completely unfair and completely one-sided. It requires a legislative remedy because employers, left to their own devices, will not engage with the industrial relations machinery of the State. Ultimately, Covalen is owned by a company that benefits from massive public sector and State contracts, so it is benefiting from taxpayer funds and it is not engaging with the industrial relations machinery of the State.

I am sorry to cut Mr. McArdle off; I am just conscious of time. We will have a second round of questions. It is 12:45 p.m., so we will probably have three minutes per member, if that is okay. I ask members to ask one question because by the time people ask a question, the time for an answer is nearly over. I ask everyone to be succinct and I remind members that we need to finish our interim report. We will have three minutes per member, starting with Deputy Mythen.

I will finish on the point Deputy Geoghegan and Mr. Reidy made about the AI summit. The Irish Nurses and Midwives Organisation, INMO, the Irish academy and musicians and artists were not invited.

The Financial Services Union in its statement said that the approach so far in Ireland's policy is "with less emphasis on socioeconomic equality and workers' participation". What active role is the FSU taking and what is it doing to try to influence the policy?

That is why Ms Newell is here today. We have done research and we are now into the second phase of research. It is some of the first research in Europe that is being done into workers' views of AI and its impact and what their fears are. As I said, we are on the second round in relation to that.

To go back to Mr. McArdle's point, and it is important to labour this point, the issue of collective bargaining is at the heart of the future of change. I would not dismiss easily the points Mr. Reidy made in terms of social cohesion and so forth. We have members in a small credit union in Tullow, County Carlow, who are on strike for recognition. They are all women, the vast majority of whom, with one exception, have joined the union. If we were on strike 60 miles up the road in a credit union in Northern Ireland, we would automatically get recognition in Northern Ireland and have done so on various occasions in terms of representing workers.

There is a deficit at the heart of employment law and this is going to expose it because it is going to be a pretty massive change. Covalen is how not to do it in terms of change. Bank of Ireland is probably an example of how it can be done, and we will see how that works. In that case, we were able to sit down with the employer and discuss the kinds of issues that are concerning our members and try to put some guardrails around that. The people in Tullow have been working in that small credit union for eight years, with one pay increase in that time, and we cannot get the employer to sit down and negotiate with us. That is the weakness in the legislation. It is for big employers like Covalen and it is for small employers. That is the big deficit that we see in this situation. We know the pace of change. We have been through digitisation, way back to when ATMs were introduced. We have done technology up and down in terms of what it entails, but we know the changes that are envisaged in relation to AI are of a scale that they require a different approach. That approach, in our view collectively, is collective bargaining.

I thank the witnesses for staying on for more questions. I mentioned young people earlier. On the radio last Saturday, in the context of a discussion on the talks happening at the moment, it was stated that a lot of entry-level workers feel they were left behind in some public service pay agreements. I am not asking the witnesses how they might address that in the agreement they are negotiating now but I wonder whether the union movement is looking at the transition to AI in terms of its impact on entry-level workers, young workers and graduate-entry workers. It is one thing to talk about how existing workers are using AI and making sure they get the benefit of the productivity gain that might arise for the employer. How is the union movement looking at the impact of AI on young people in terms of a potentially reduced number of graduate jobs being available or, alternatively, future jobs that might be available?

Research studies show that young developers are quick to use AI and know how to use it. However, in order to arrive at a complex solution, the research shows they do not have time to develop all the methods to solve the issue themselves. Older software developers, on the other hand, also use AI but more as a tool because they know how to solve complex issues. Younger workers will know how to use AI, but AI will, at some point, lead to a wall. If we do not leave the space for young workers to learn the methods and the depth of the work, we will have another issue with those young people. It is "Yes" to the use of AI but they need to take the time to learn their work and that is a question for organisations.

I thank Dr. Cefaliello. The Chief Information Officer, appearing before the committee some months ago, identified the areas of the public service being looked at as part of the 2030 digital strategy. I understand that has now been narrowed down to four areas. If the introduction of AI in certain areas of the public service ultimately results in there needing to be fewer human beings working in those areas, is that automatically viewed by the unions as a net negative or do they stand back and observe that efficiencies are being delivered in one part of the public service but we will always need more nurses, SNAs and teachers, none of whom will be replaced by any productivity gain? Is there that type of holistic view within the movement?

That is a key question and a perfect example of why we need to have a proper collective bargaining engagement that looks at the big picture. Clearly, we cannot leave all the burden of work on nurses and teachers. The question is how to share the burden. One of the things we need to look at, for example, is introducing a shorter working week. That has been piloted globally and has been shown to be successful in terms of the productivity it delivers. We need to sit back and look at all of this in the round. We deliberately referenced 1987 today in terms of the ambition we need when discussing AI in public services. How do we get the best out of it while also ensuring the benefits are shared across work and across public servants? I cannot answer the Deputy's question today because we need to find the answers through that engagement. Collectively, we can find those answers.

Is the funding of reskilling-----

You are a minute over your time, Deputy Geoghegan.

May I have just a few seconds more, a Chathaoirligh?

No. I am sorry but I must allow in other speakers. Deputy O'Rourke is next.

My questions are along the same lines. A huge part of what the witnesses have said is that, ultimately, there must be a framework for engagement on this process of change. What would sharing the productivity gains look like? Is it, for instance, investment in retraining and reskilling, a shorter working week or better terms? Will the witnesses set out how that might look? I think I know the answer to this but is the union side arguing that this will be achieved through collective bargaining and system-wide engagement?

Is it the case that the union side argues it is achieved through collective bargaining and system-wide engagement?

I will respond briefly and then hand over to colleagues. There should be a three-layered approach to this. There should obviously be some engagement at international level, and a lot of the people at the AI summit said we need global regulation, but there needs to be a shared understanding between employers, unions and the Government around AI, so a framework and vision. Then there needs to be sectoral task forces where specific sectors can come together, both on the employer and union sides, with the appropriate support, to try to see how we can look at this huge factor of change and how it can work for both sides of the labour market. There is then obviously a different agenda but similar discussion at enterprise level. It has been done before with regard to gain-sharing and world-class manufacturing back at the change of the century when different models of work were introduced. The Deputy summed it up: sharing should be better wages and training, and more free time. We are seeing more and more wealth going to capital and less going to labour, to use that terminology. We need to address that or else we will have greater inequality being stored up and the social contract being breached for young people here.

I will make a quick point on how one might share the dividend. Burnout for female workers was referenced earlier in the discussion with regard to the domestic pressures they have with caring responsibilities. Our ageing population was also referenced. One of the problems the Government faces is how it supports carers. An awful lot of the care of others is done by women who are also working. Part of the social cohesion and positive development of this changing world of work, to which my colleague Mr. Reidy referred, would be to look at how we support people in the workforce who also have caring responsibilities. We have had questions about the data, money and finances, but we need to look at ourselves as a country and society and how we can use this shift to make things better for everyone.

I am going to defend the international AI summit because the focus of the summit, and very clearly how it was presented, was to revolutionise Europe's competitiveness. On the same day, there was a summit taking place around copyright and intellectual property, IP. It was open to all members of this committee to attend both of those. The Government is taking all of these issues seriously. Whether it is a case of trying to bite off all aspects of IP or AI at the same time, it is difficult.

On the issue of the creative industries, we were meant to be dealing with the module on the creative industries right now but, against my recommendation, the committee chose to bump it to next year. I want to defend the AI summit. It is important that we continue to attract FDI because it is critical to fund the economy, but there are also a number of trade union members working in those companies. It comes back to this question of how we are going to sustain livelihoods because the concept of the job for life is gone, and it is about making sure we are working alongside AI in safe and responsible spaces.

I want to look into the future when some of the debates about whether we need to look at guaranteeing a basic income for all workers in the State are going to happen. In other words, how do we ensure we sustain livelihoods alongside these transitions to technology? Technology is going to be doing things that we now take for granted. Our challenge is how to equip every citizen. It is about avoiding a digital divide that could emerge. I do not know if the witnesses have looked at a position in that regard. I appreciate their focus is on pay talks but we could start to look at the world of work in 2030 and 2035.

There were some very interesting points there. Regarding defending the AI summit because the focus was competitiveness, competitiveness is a relative concept. What matters for economic growth and people's well-being in the longer term is actually productivity. Part of productivity is having people be part of the conversation, learning from workers, and horizontal innovation. It is all of these things. Excluding workers from that process is missing a trick, frankly.

It will also enable companies to adopt and diffuse more easily if they have workers on board and things are designed in a way that is pro-workers. The Deputy made a very important point about how we are going to sustain livelihoods. Technological change is happening all of the time. There is a lot of hype about AI at the moment. Some of it is justified and some of those hypemongers like to focus on issues of basic income. They do so partially because it says people do not need to worry about work or about its transformation or about inequality. First, on basic income, a lot of a person's well-being comes from the work that they do - feeling needed, having agency and being respected and valued. Moving to a system where people are just living on basic income, for example, if that is where we go to, is a recipe for disaster. It is the opposite of social cohesion. People need to feel valued and moving to a basic income context would be a dystopia, frankly, if that were to replace work.

Indeed, on basic income as a concept, if basic income is to be sufficiently large enough to be meaningful, let us say it was €20,000 to €30,000 per person per year, the sums add up very quickly. In terms of the scale of the Government's welfare budget, it would be orders of magnitude on top of where it is now in a context for just fewer people working. Then, there are questions about whether we are just simply going to tax capital to try to rebalance it. What we need to think about with AI infrastructure as well is who controls the AI infrastructure. Europe absolutely needs that as well, but we need to think about moving towards a utility model and treating it as a utility rather than just something that can be privately owned with the power that comes with that.

I put on record my absolute support for collective bargaining. As a student of industrial relations under the former trade union leader, Joe Wallace, in University of Limerick, I fully appreciate the absolutely vital role that unions and worker organisations have in our economy and the contribution they have made to the success and growth of our economy through many decades. I am curious. I apologise, because I was in the car for the very early part of everyone's opening statements so I do not know if they touched on some of this outside of what they had sent in advance. I will ask, organisation by organisation, whether they have positions on data centres and energy consumption by data centres. They may or may not have an organisation position on it. I will go first to the Financial Services Union.

No, we do not have a formal position in relation to it.

What about Fórsa?

We are concerned about the impact of data centres, obviously, on the environment and on energy availability. We want to avoid workers having a double hit on the cost of energy and the cost of living and then access to actually having that energy.

The Irish Congress of Trade Unions?

We do not have a formal position on it but I do not think they have made the case for themselves as good as they should. It is not yet compelling.

I agree. What about the European Trade Union Institute?

We do not have a position. It is part of the cover when we do just transition. Basically, it has been set upon the question of workers extracting lithium and all of this, so we look at the invisible part of AI. There is research being done, but not a position.

I thank Dr. Cefaliello. What about the Communication Workers' Union?

We do not have a formal position other than to say that we recognise the fact that the recent report that claimed there were 830,000 jobs dependent on the data centre infrastructure was found to be kind of misleading, really. There are 3,300 direct jobs that depend on the data centres. We represent data centre workers. I should say that up front. We would share concerns. We have worked with Friends of the Earth and we would be very concerned about the growth in electricity demand. It is up to about 23% now for data centres, which outsizes the demand generally in the sector.

I thank everyone for those responses. For me, it is impossible to think about the speed and pace of AI adoption, the energy consumption, the database requirement and our need to satisfy that and decarbonise it. I am very focused on the potential for floating offshore wind development off our west coast, which would take some of the heat out of the onshore solar or onshore wind development that is happening around the country. That is another challenge to societal cohesion, particularly in rural areas. I hear that weekly from our colleagues in our parliamentary party meetings. It interesting to know the representative organisations' positions. If we are moving into this world where energy technology and digitisation is increasingly important and a component of work, will they also be moving into a space and taking positions on responsible generation of energy and management of that data?

I thank the Senator. Everyone will be happy to know I have the last two minutes for questions. The Financial Services Union spoke about AI adoption and workforce preparedness. Deputy Geoghegan asked whether people are using AI within the unions and Mr. Reidy, Dr. Cefaliello and Mr. McKinney spoke about different instances as well.

How AI is being predominantly used right now, and people can agree or disagree with me, is really basic. It is basic administrative tasks. It is helping to compile reports or trying to digest reports or helping people with some sort of campaign or to come up with a few slogans for a campaign. It is pretty simple. Yes, there is agentic AI. There are different examples in different businesses where they have developed more profound artificial intelligence. Where we are at the moment, I do not think that type of AI is replacing people's jobs and I do not think it will because there is no critical thinking. It still needs prompt engineering and the outputs are fairly simple, but they are good and useful on an individual basis. We are at a critical moment and the witnesses have all kind of pointed at this because, as AI develops and new types of agents are created, where it could displace a job or most of those roles and where I think we need to be particularly careful is with the more administrative roles. That is why I drafted and created the national AI workforce transition programme document. For those who have not received it, I will share it. That type of policy is where we should be moving rather than people being fearful that it is going to take their jobs next week. I do not think that is happening. I will direct this to the FSU because I referenced it. Is there is there any particular kind of policy shift we should be looking at or does Mr. O'Connell think we should be moving in another direction?

I recall the paper the Cathaoirleach is speaking to and I think that is part of the direction we need to travel in terms of upskilling. Again, however, it has to be different than before. A lot of the upskilling that takes place in our sectors is where people have to carve out the space themselves. Reskilling of this nature requires protected time and investment in people. That is what the paper was speaking to and it provided evidence from Singapore and some of the Nordic countries where a specific approach was taken because of investing in people. That is part of the dividend for people in terms of the future and future-proofing their roles and upskilling them. There are other elements of it where sharing of benefits and so forth are part of it as well. People like to have a say in what their role is, what the role is transitioning to and what it is that comes easily to them or perhaps does not come as easily to them and they need upskilling.

As an organisation, we were asked the question earlier in terms of investment in our staff. We invest in our staff, everybody from PhDs down to higher diplomas in various areas and skills. We have invested in AI. We came from a staff meeting this morning where we were talking about technology because we are implementing a new customer relationship management, CRM, system and so forth. It is, therefore, part of our daily interaction and our understanding, but there are roles. We represent animators. They could be literally wiped out in terms of AI because of what it can do and what it is capable of, so that would be a skill completely lost. Risk management is another area where it is identified. Then, we can extend it to journalists and other areas with the law and so forth.

I thank everybody for coming today. It has been extremely worthwhile and extremely eye-opening for the committee to hear all of the witnesses' viewpoints and consider them. It is much appreciated. As I mentioned earlier, I appreciate that there is a lot happening in the world of trade unions at the moment, so we appreciate people giving us their time today. We will continue with AI work and, after this, we will be looking at AI in the creative industry. That will be a continuing theme.

I propose to members that the meeting be suspended for ten minutes for comfort breaks, etc. We will come back then in private session to go through our interim report on AI truth and democracy.