I welcome everybody to this morning's session of the Joint Committee on Artificial Intelligence. I particularly welcome the transition year students who are with us, and the other visitors in the Gallery. They are all very welcome to our discussion. The committee comprises nine TDs and five Senators. It considers how the State should respond to the challenges and opportunities of artificial intelligence. The committee operates in a module format, so issue by issue. One of our earliest sessions was on the general theme of AI, children and young people. We are now looking at questions around AI, the State and democracy. In this morning's session, we will discuss with a number of people with an interest in media and journalism the impact of AI in that area.

We have received an apology from Senator Dee Ryan. Deputy Naoise Ó Cearúil is at the comhcoiste na Gaeilge and will join us shortly. Other members are on their way.

I advise members of the constitutional requirement that they have to be physically present. I am quite certain our guests will be very aware of the constitutional requirements on witnesses before an Oireachtas committee.

We have quite a number of guests today. From the National Union of Journalists, NUJ, I welcome Mr. Séamus Dooley, assistant general secretary, and Mr. Gerard Cunningham, chair of the NUJ Dublin freelance branch and treasurer of the NUJ's Irish executive council. From Media Literacy Ireland, I welcome Dr. Eileen Culloty, co-chair of Media Literacy Ireland and co-ordinator of the European Digital Media Observatory, and Ms Martina Chapman, national co-ordinator of Media Literacy Ireland and independent consultant on media literacy. From NewsBrands Ireland, I welcome Ms Sammi Bourke, chair of NewsBrands Ireland and operations director with the Irish Farmers' Journal, and Ms. Deirdre Veldon, vice chair of NewsBrands Ireland and principal adviser on media and policy at The Irish Times Group.

I thank all of our witnesses for being here today. I invite each of them to deliver a short opening statement. We will start with Mr. Séamus Dooley from the NUJ.

I thank the Cathaoirleach, Deputies and Senators. The NUJ is grateful for the committee's invitation to discuss the theme of AI, truth and democracy. The collection, verification, crafting and dissemination of news are not merely an economic activity but are of fundamental importance in a healthy democracy. There are very good reasons those who seek to abuse power so often begin by undermining press freedom.

The NUJ recognises the potential benefit use of AI as an assistive tool, with human oversight and control. We also believe that at both national and EU level, there is an urgent need for oversight to block monopolistic tech firms with deep pockets from dictating the shape and content of information and the working conditions of those who craft news. Even since I typed that paragraph, the reality of that has been reinforced by the devastating news about The Washington Post, where so many jobs have been abolished. This includes all photographers in the same week in which Jeff Bezos invested $200 million in AI. What does that suggest about his vision of the use of AI and visual imagery?

AI is not an impending threat but a presence across our industry. Threats of exploitation and intellectual copyright breaches have already been realised, with false attributions, inaccurate AI-generated stories and creators discovering use of their work without consent or compensation. Such abuse of AI erodes public trust in journalism and journalists and risks the reputation of every journalist and media organisation seeking to adhere to ethical standards.

The interim report of this committee contains 85 recommendations but is remarkably silent on the issues of employee engagement, trade union consultation or collective bargaining in relation to the major transformational changes across a range of industries.

On 17 December 2025, the European Parliament voted in favour of EU actions on digitalisation, artificial intelligence and algorithm management at work. We support the call by the European Trade Union Confederation, ETUC, for a quality jobs Act and a dedicated EU directive on AI and algorithmic systems in the workplace, including a range of measures. By definition, AI is designed to limit and centralise human intervention. In that sense, it may be said to run counter to democratic principles. In some respects, talking about AI and democracy is a contradiction in terms.

AI influences the shape of news, the manner in which it is processed and transmitted, how it is understood and, crucially, how it can be manipulated. The threats and challenges posed by AI cannot be viewed in isolation from the vulnerable position faced by workers denied the right to collective representation under legislation in this country. Contrast that with the process of engagement and negotiation which took place in our industry during what was called "direct input negotiations" and the introduction of various technological phases in the Irish print media.

In the media sector, the inadequate response by legislators and regulatory authorities to the concentration of ownership poses an additional threat to diversity and plurality and is not separate from the topic of democracy. In the face of rapid redeployment, the Government must introduce comprehensive legislation regulating AI while enforcing existing copyright and intellectual property laws. Creators’ economic rights must be protected from exploitation, with journalists having full control over their work, how it is used and abused and what they are paid for it. The NUJ calls for a simple opt-in mechanism, either for individuals or as part of collecting licensing, revocable by the creator at any time. Fair and reasonable terms of compensation must be secured for all journalists and creators whose work has been used without consent.

Transparent labelling is key to maintaining public trust and quality standards. We also believe the AI office must evaluate changes in employment level, job tasks, hours worked, wages, fees and royalties in creative occupations associated with generative AI adoption, including both increases through productivity enhancing augmentation and decreases due to task of job substitution. In any operation of the AI office or in any further discussion, there has to be consultation with the media industry which has been excluded to date in the State's formation of policy. We agree with the NewsBrands position on that, which in itself is a rare enough occurrence.

I thank Mr. Dooley. Our interim report before Christmas dealt with the issues the committee had dealt with until that point. I have no doubt we will make further recommendations.

I now invite Dr. Eileen Culloty to give the opening statement on behalf of the European Digital Media Observatory, EDMO.

I am speaking on behalf of EDMO Ireland, which is an EU-funded project to address disinformation through research, fact-checking, media literacy and policy analysis. We work closely with Media Literacy Ireland and share its view that AI literacy is essential. Based on our work on disinformation and platform co-regulation, EDMO Ireland has specific concerns about AI’s implications for truth and democracy. Here, it is important to distinguish between AI as research, AI as tools and AI as a new frontier for major tech companies. Our concerns lie with how this last category is reshaping information environments, not with the legitimate use of AI in research and applications.

Regarding truth, generative AI is fundamentally unreliable. It does not aim at factual accuracy. Instead, it offers outputs that are statistically plausible based on its training data. This limitation is well known. Nevertheless, tech companies overhype AI systems and embed them into everyday life. We are offered convenient, but unreliable, tools to generate, summarise and evaluate content. Evidence from the tech companies themselves suggests a dependency on these unreliable tools undermines our human capacity to reason.

In summary, the problem is not only that generative AI is flooding the world with scams and low-quality slop, which it is. Nor is the problem simply that technology platforms’ own ability to detect and label AI-generated material is wholly inadequate, which it also is.

The deeper issue is that flawed and opaque AI systems are undermining human reasoning and eroding a cornerstone of democratic citizenship and public debate.

In the 1990s, the policy approach to emerging technology companies was defined by weak regulation to support innovation. We are still dealing with the consequences of that. Today, those same companies are global giants, many of which are hyping a new technological revolution. Faced with regulatory choices, it is striking that we appear to be making the same mistake again.

I thank the Cathaoirleach and members of the committee for the invitation to contribute today on behalf of Media Literacy Ireland.

Artificial intelligence applications and technologies are expanding at extraordinary speed and are being proposed, and in some cases adopted, across education, health, defence, public administration, and employment. Advocates emphasise innovation and efficiency. Critics point to fundamental technical errors, opaque systems, environmental costs and the commercial and political motivations of the companies deploying these technologies at scale.

These are democratic questions that affect all citizens because significant consequences of AI are felt at social, political and democratic levels, as well as on an individual or personal basis. For people to engage with these debates about AI policy and for people to navigate AI in everyday life, people must develop a foundational level of knowledge about AI. This may include knowledge of ethical literacy, data literacy, scientific literacy and design literacy.

As a cross-sector national network promoting a life-long approach to developing media literacy skills, Media Literacy Ireland is well placed to help equip people with the skills and knowledge required to make the best use of evolving technology such as AI. It is also well placed to address the intersection of AI with existing issues addressed by media literacy such as disinformation, evaluating trustworthy information and democratic participation.

Media Literacy Ireland maintains that life-long media literacy, for all citizens, remains essential not as a substitute for regulation or accountability but as a means of public empowerment. Moreover, as with other topics addressed by media literacy, caution should be exercised to avoid placing too much responsibility on the individual for the safe and ethical use of AI while downplaying the obligations of companies and policymakers.

Important work is already underway, from organisations such as the ADAPT centre and EDMO Ireland. However, a more co-ordinated approach would help to ensure that all citizens have access to foundational AI literacy as part of their life-long learning journey in order to help them make informed choices about whether and how they want to engage with these technologies.

I thank the Chair and members of the committee for the opportunity to address them on behalf of NewsBrands Ireland, the representative body for Ireland’s national news publishers. This discussion is timely because the actions and decisions taken now will determine whether AI strengthens our democratic institutions or quietly destroys them.

NewsBrands Ireland represents publishers committed to a free and independent press, high editorial standards and a sustainable news ecosystem. Our publishers invest heavily in professional, fact-checked, multiplatform journalism that builds their audiences’ trust, serves communities across Ireland and supports the wider media ecosystem. However, our businesses have been undermined by the big tech platforms, AI and changes in audience habits.

Publishers in Ireland are now under grave threat. As a result, so too is truth and so too is democracy. It is worth making a distinction between news and journalism. News is just that - new information about recent events - whereas journalism is the process of gathering, verifying, contextualising and presenting news to the public. If our media businesses fail, news will still happen, but independent journalism in Ireland will no longer exist.

Globally, publishers are sounding the alarm about the unchecked use of copyrighted material in AI training. Generative AI is powered by journalism it refuses to pay for. By harvesting original reporting to feed large language models, LLMs, the financial viability of quality journalism is under existential threat. If journalism becomes unsustainable, the consequences will be weaker scrutiny of power, greater exposure to disinformation and misinformation and declining public trust in democratic institutions.

News publishers are not anti-innovation or anti-AI. We have embraced technology and we are deploying AI tools as possible within our businesses. Progress must be lawful, transparent and accountable.

The current programme for Government commits to a strong, independent media sector and to insisting on payment for the use of content. Those commitments now need delivery. We are calling for robust Government action to ensure AI systems operating in Ireland are transparent, accountable and compliant with copyright and data protection law. Technology companies must disclose what content they use and how they use it. Fair remuneration for creators is not optional; it is the price of a sustainable information ecosystem. Notably, Ireland’s own national AI advisory council has acknowledged that our copyright framework is under strain from rapid AI development and that gaps and ambiguities, particularly around reproduction rights and text and data mining, threaten Ireland’s creative economy. It is remarkable that no input was sought from NewsBrands Ireland, nor it seems from Ireland’s media and creative sectors, by the AI advisory council in formulating its advice paper, The Impact of AI on Ireland’s Creative Sector. Maybe this is not surprising, though, when you consider that the council's membership, which includes representatives from AI companies, has no representation from the media or the arts.

The EU AI Act provides critical safeguards for creators and rights-holders. Chief among them are the requirements that AI companies publish detailed summaries of copyrighted data used to train their models, which should allow publishers to be compensated for its use, and for the clear labelling of AI-generated content. Ireland can be the leader, particularly during its Presidency, in ensuring the provisions of the AI Act are implemented in Irish law and in calling for reviews of other regulations from the EU to ensure they are still fit for purpose.

Is the Government doing enough to support media outlets in extracting their due from big tech companies and now AI platforms? Is Ireland over-servicing the needs of big tech at the price of the media’s future? The publishing industry believes it has little support from the Government to assert its ownership of original content. We are living through the AI gold rush and Ireland has a vital opportunity to help in shaping the information marketplace and to sustain responsible, accountable journalism while so doing.

As responsible publishers, and through our NewsBrands Ireland programmes, we already play a role in helping the public understand the difference between fact-checked, trusted journalism and misinformation and disinformation, and in highlighting the dangers of AI-manipulated and some social media content, particularly for children. Ireland’s EU Presidency presents a strategic opportunity to lead Europe in building AI governance that protects democracy, safeguards creativity and sustains independent journalism. NewsBrands Ireland is committed to working with the Government, regulators and industry to help counter misinformation and deepfakes, to improve media literacy and to secure publishers’ rights and the future of quality journalism in Ireland. The time to act is now. I thank the committee.

I thank Ms Bourke very much. We are now going to go to questions and answers from members. They will have seven minutes for questions and answers, so we ask people to be succinct in their questioning. I call Deputy Darren O'Rourke.

I thank the witnesses for their opening statements and the additional information they have provided. I will start with a question related to media literacy and AI literacy. I put this question to Dr. Culloty and Ms Chapman. Can they reflect on the current regime? What are the immediate opportunities for improvement? I am a spokesperson on education and a lot of the time Finland is pointed to as an example of a system that has embedded social media and AI literacy from a very young age. Are there immediate priorities from Ms Culloty's and Ms Chapman's perspective?

On education, AI literacy is quite new for teachers. With media literacy in general, teachers have gone through their teacher education and have not been given any instruction on media literacy or AI literacy. The ADAPT centre, funded by Google, has been doing AI literacy workshops with teachers around the country.

What is very striking when you meet teachers is that they are already locked into AI platforms. They are Google schools or Microsoft schools. It is already there. The young people in their care are already using them, and they feel like they are in an uphill battle. As Ms Chapman's statement noted, everybody needs a basic level of AI literacy to now comprehend what is going on around us, but it is a mistake to think AI literacy in itself can help address these issues.

If I may come in there, beyond the formal education system, we need to be really careful because AI literacy, that is, the need to understand and to be able to make informed choices about how we use AI, affects everybody. There is rightly a focus-----

On schools.

-----on supporting young people, and through schools is the obvious way, but we cannot not look at the bigger picture and wider society as well. That is why the approach Media Literacy Ireland takes in terms of the broader media literacy piece is working through trusted third party intermediaries to help reach people where they are.

What does a governance structure or a regulatory structure look like to deal with the breadth of that? Does it fall within individual Departments? Is it the national AI office? Is it all of it?

Certainly, and this is not unique to Ireland but applies everywhere, because media literacy touches on so many different topics that sit within different policy areas, there is usually no one single Department or one single body that can deliver media literacy in isolation. In Ireland, Coimisiún na Meán has a statutory duty in relation to media literacy, and there are many levers it uses to discharge that duty. One of them is supporting Media Literacy Ireland, and that is how the cross-sector approach is delivered.

Based on where we are at the minute, would the witnesses say there is a need within that ecosystem for additional resources, support, funding or capacity or are we missing something entirely?

This is a growth area. Every time technologies change - sometimes there are even changes in legislation or changes in social norms - it brings a situation where we as citizens need to develop new skills, so there will always be a demand there and it will always be a growth area. It is very difficult to predict the future. Also, everybody is on a very individual learning journey, so the skills I have today might work, but tomorrow I might need a different set of skills. The key - and I am not sure we have this cracked yet - is that it is about a behavioural change with people. It is about trying to get everybody to a point where they understand that they need to constantly update their skills. We are all busy people, and the last thing we want to do is sit down and figure out, "Today what will I learn about or what skills do I need to develop?"

May I just add, on AI literacy, that it is really important to note that there is quite a lot of tension and debate around how AI literacy is defined. Many academics are very critical of the way AI literacy is being proposed as a responsibility of individuals, that they must take on the burden of educating themselves. Even major bodies like the OECD have come out with definitions of AI literacy that are very heavy on the personal responsibility. When we talk about frameworks, there is a lot of work just to be done to define what exactly it is that we need out of AI literacy.

I thank Dr. Culloty for that.

Mr. Dooley might want to pick up on some of that, but I want to ask him whether there are specific areas of journalism that are most at threat. He mentioned photography, photojournalism and The Washington Post as examples.

I ask him to comment on the concentration in the media, the small number that have-----

First of all, I agree with the notion that the onus should not be on the public to read misinformation or to understand how they are not caught out by AI tech giants. The onus should be on legislators to ensure that they end the love affair with tech companies and challenge them on their misinformation. That would be a good starting point.

I would say there are very few areas of journalism not touched by AI but photography is a real concern. Last Friday night we had an amazing celebration of press photography in Ireland, as we do every year. The reality, however, is that the misrepresentation of photographs and the use of visual images in all of this are a major concern for democracy.

The issue of copyright and the issue of the labourer being worthy of their hire are relevant to AI. The wholesale robbery of material owned by our members is of concern to us. We do not always agree with publishers - I am not naming anyone in particular - because sometimes they do not recognise creative rights. There is always a tension about creative rights, but certainly work should be paid for.

Concentration of ownership is a huge issue in Ireland. The Chair will probably agree that is an issue more for the media committee, but it is relevant to this committee as well. One of the big concerns we have is that, while we would support a just transition fund in relation to this, Coimisiún na Meán is doling out money left, right and centre under democracy schemes to employers, particularly one individual in the regional press sector, that refuse to go to the WRC and refuse to recognise trade unions. There is no right for workers in that situation to deal with that. That is why I am saying that you cannot just take AI as one issue in itself. On that point, I believe there is a need for State assistance but that State assistance must be linked to procurement rules.

I thank everyone for their presentations. I will make two comments first and then I have a couple of questions. I acknowledge the stuff with The Washington Post, given the billions of euro that was promised when it was taken over. A relative of mine is a visual journalist with The Washington Post and I went from seeing images of him working with tear gas in his face to try to capture what was happening in Minnesota to, within 24 hours, hearing he was not able to access his Washington Post email. To watch that transition and how quickly it happened without warning is extremely scary in terms of truth and democracy as to how we talk about it today.

Deputy O'Rourke talked about AI literacy. I acknowledge Dr. Culloty, who did a great piece of work in the Irish Prison Service alongside me and the OSCE in relation to media literacy. We put it together very quickly and it was responded to extremely well. Often when we think of how we do literacy and media literacy, or AI literacy or whatever layers we need to build onto it, it is a question of how you get into those communities who will be the furthest from being able to access any sort of learning or literacy in the normal ways. It is on us, communities and community workers to try to find ways to bridge the expertise that exists within the witnesses' sector but also community development and how we can embed it in community development practices in order that we are developing accessible literacy initiatives that are not too top heavy or too academic heavy.

I have two questions maybe for Dr. Culloty and then one more general one. I might ask all my questions first, because there are three, and just let the witnesses come back to save me going back and forth. I am wondering about the content labels with regard to disinformation or deepfakes. Can the witnesses maybe talk to that a little? Is that enough? Is it not going far enough? What way will that help or not help?

In relation to the EU Act, do the witnesses have any comments on the risk-based approach as it relates to the DSA specifically?

My final question, which is maybe for Mr. Dooley and Ms Bourke and their colleagues, is more about the European Commission and Google. Obviously, we know that the websites of media outlets and newspapers use Google in relation to the ad auctions. Do the witnesses believe, given the illegal monopoly that Google has in this regard, that media outlets - newspapers or whatever - should sever their relationships? Mr. Dooley spoke about ending the love affair with big tech, but then we look at ads and Google ads and how they are embedded. What do the witnesses think of that, and should the European Commission act against Google's unfair anticompetitive business model?

They are my questions.

A lot of big questions.

On the content labels, the EU is developing a code of practice on labelling AI-generated content. The EU is very good at coming up with these kinds of codes of practice which say platforms should do X, Y and Z, and then a bunch of people, often academics or civil society actors, have to go and read those reports.

It is easy to put those things in place. Whether they are effective in practice is a different question, however. The issue with content labels is that we have an in-built bias to assume that transparency is a good thing that will have good outcomes, but there is little evidence for it in this case. Putting a label on AI-generated content is not the same as putting a label on, say, tobacco or alcohol products, which are bounded things. AI-generated content could be about anything. Some of it is fully generated by AI, while some is manipulative deepfakes. Some of it is also people working with AI for their creative production. There is a wide range of stuff in that regard.

It also puts an onus on creators to participate in good faith labelling. If someone is intent on manipulating a democratic system, however, why would they bother to do that? There is also an assumption - an assumption they promote - that the technological platforms have the capacity to detect content, but there is no good evidence that they are good at detecting content proactively. There is also the fact that if we have a lot of labels, people become blind to them and ignore them. There is a lack of good evidence that it is effective. During the presidential election, there was a deepfake of Catherine Connolly, along with other deepfakes impersonating politicians and making up fake things. They were left online even after they were reported. They never had labels put on them, and that is during the presidential election when one would expect platforms to be proactive.

In the context of the risk assessments, comments from some of the previous people who came before this committee to talk about, say, AI in the public sector have been interesting. Conducting risk assessments on the AI tools that will be deployed makes perfect sense in contexts such as the Department of Social Protection or the Citizens Information Board. When it comes to risk assessments for OpenAI, however, who is auditing OpenAI’s risks assessments? It is not clear that this is ever going to happen. Risk assessments are built into the Digital Services Act, DSA, and look at the crisis that we have had already this year where actual risk assessments are in place. Risk assessment is good in theory but there is an awful lot for that to actually make sense. The reality is that the big AI companies are ploughing ahead with an accelerationist view. Their approach is to move fast and breaks things. I fear that our risk assessments will become mere compliance reports.

I ask the witnesses to speak to the EU Commission piece.

The Senator talked about the ad market in Google and what is happening in that regard. In the context of Google AI Overview, referrals coming to businesses, particularly from publishers, are accounted for. When using Google AI Overview, people are not clicking through. People are not traditionally checking through to determine where that source has come from. Previously, publishers were dependent on referral traffic to their website via Google searches, but that no longer happens. In recent research that came out in December, a total of 76% of all users who use any type of AI overview are not clicking through to the source. It is not just the ad market itself; it is also AI from a content perspective driving it back to the source, that is, the publishers. That is a big concern for us as an industry.

Google, Meta and X have, metaphorically, got away with murder. They have insisted, and it has been bought into and accepted, that they are merely platforms rather than publishers. That is a core problem. No one else gets away with saying that they only lit the match. The fact of the matter is that they set the place on fire and did a huge amount of damage. The platforms have pillaged and profited from the media sector with financial immunity. I am not sure whether there is the power to deal with them by saying we are going to outlaw them. We have long-called for a windfall tax on tech giants because the only thing that capital understands is cost. A windfall tax on practices that involve regurgitating the media industry across Europe seems to me to be more effective than talking about trying to ban them.

I thank all the witnesses for coming before the committee. I will start with Ms Bourke and Ms Veldon. Ms Bourke mentioned the interesting figure that approximately 70% of those using Google AI Overview do not follow through to the websites. Witnesses from the Citizens Information Board attended an earlier committee hearing and they highlighted that its website has seen a traffic collapse, essentially.

They were concerned that people will read the AI summary, which might take bits from the Citizens Information Board and other bits from various articles that might be outdated and cobble them together, leaving readers with 50% to 60% of the information being correct and the other 40% incorrect. They are trying to work out how people can discern the middle of all of that. How will news organisations, like those Ms Bourke and Ms Veldon represent, survive in the AI summary world if traffic is not being directed towards websites? What is the future model? Will premium, paid models be sufficient to act as a wall to AI summaries? What does the future look like?

It is a difficult one. All publishers are looking at their business models regularly to see how they combat that. Subscriptions and looking at premium subscriptions are definitely one way. The issue we still have is that research has been carried out that shows that content behind pay walls is still being scraped. Protecting our unique content and the work we are doing is still difficult. Subscriptions are the way. Transforming the way in which we operate is what we are looking to do at the moment, but it is still not going to be enough in the future.

To add to that, if people look at what AI mode, AI overviews and ChatGPT can tell them this morning about live events in the UK, they will understand that this scraping that is going on is happening live before our very eyes with our content. To go back to Senator Ruane’s point about whether we should break all ties, in the same way that we did not ask for our content to be scraped by search engines but we later became dependent on them, we did not ask for our content to be scraped by AI platforms. That is exactly what is happening now and they appear to have full deniability about that, which is a remarkable thing to see.

The reality is that publishers do not have the answer as to how we overcome this. We have a dependency on search. The level of direct traffic in countries like Ireland is quite low overall compared with Scandinavia, for example, where the English language is a particular problem there. We do not have the answer. We are working as hard as we can to develop original content that AI cannot replicate but we need help to do that. We need innovation funding to allow us to develop models. In the time to come, we have the spectre of liquid AI, for example - it is with us at the moment, in fact – and agentic AI where people will have all of their questions answered, possibly in a single ecosystem. In that context, the worldwide web, as it is now, will cease to exist and publishers will be in no way able to replicate a single news ecosystem in the way that a massive AI platform could now do.

Ms Veldon mentioned organic content that AI cannot replicate. What does that look like?

It is original content for the most part. A lot it is around podcasts, which are quite hard to replicate. Synthetic voice is still a poor replacement for the human voice. Other original content includes investigative work, particularly deep-dive journalism, content around colour, and information on what is behind the news. We are also making an effort to help people understand news processes and how we make journalism, including the checks we go through. We think there is a literacy piece when it comes to getting information across in an interesting way that helps people engage with our content.

Does Ms Veldon think, in this brave new world that news organisations have to function under as private operators, that it make news more elitist if core news can only be accessed through high-level premium subscriptions? In some ways, people can argue that AI is a bit of a leveller, or it can be a leveller in other areas. Digital transformation in a lot of areas of life has been a bit of a leveller. In the news industry, however, could it serve the opposite effect? Essentially, people who can pay enough to get access to those core aspects may be able to avoid the basic, general news.

We have always made it our business to make sure news gets out there. That is certainly one of our fundamental principles. We make a distinction here, as we did in our opening statement, between news and journalism.

News will happen. People will find out about things that happen in various different ways and through various different prisms. Our point is that if people want to engage with quality journalism, a support needs to be provided for that. There is a cost to be paid for that because of the level of verification and all of the different processes that are involved in producing that journalism. It is very expensive to produce, and there will be a cost attached to that. It will come down to the difference between what people get as AI slop or news in their social media feeds or quality journalism.

I thank Ms Veldon very much. I have only got a minute left but hopefully I might get a chance in the second round to hear from the other witnesses. Dr. Culloty mentioned AI literacy that she has some involvement with that is taking place in schools with Google. Could she outline to the committee what that programme is? She said something very interesting that fits with my own feedback too, which is that the kids are way ahead in some respects in how the classrooms are operating in reality. Deputy Byrne and I were in Estonia, where ChatGPT and Gemini have already been licensed in secondary schools. They are trying to fix the mismatch between what the kids are already doing versus what the teachers know. Is that the kind of programme Dr. Culloty is operating with Google? What are her thoughts on that?

No, we are doing the complete opposite. We are telling the teachers that they can take tech out of the classroom. There is a big concern from an educational point of view. The tech companies are quite open about this. They want people to become dependent on the AI summary so that they never go to the source and it replaces their own thinking and investigation. It is a very old tactic You lock people in as early as possible. The best way to lock people in is via the schools when they are young. That particular project was set up by the ADAPT centre. It got some funding from Google to provide AI literacy to teachers on the basis that teachers often just lack confidence because they have no background training in it. We have been doing workshops with teachers around the country, and we set up a course on AI literacy and ethics literacy. A common concern from teachers is that they just want to do what is best for the young people in their classroom. I can see young people here today. If teaching young people about AI is important to their future and be able to get jobs, they want to do that, but they also want to make sure that they are rounded citizens and live fulfilled lives. They are very worried about the mental health impacts of people becoming dependent on AI. The AI in education piece is a really massive one that ties into a lot of these other things. Our democracy fundamentally needs people to be able to think and reason for themselves. If we become totally dependent on a tool to do our work for us, there is no learning or thinking happening.

I thank Dr. Culloty very much.

I have to say to Deirdre that I have heard some synthetic voices that are more interesting than some of the podcasts I have had to endure. Next, I call Senator Harmon.

I thank all the witnesses who are here this morning. Listening to them, one of the things that really struck me is the fact that our democratic institutions are potentially under threat as a result of advancing AI. Something that was said, which is very worrying, is the effect on our human capacity to reason, and the effect generative AI has on that, including the fall-out that could have on democracy. As a society, it leaves us more susceptible to propaganda. This is really worrying for politics and society in general in terms of greater exposure to misinformation and disinformation.

My first question is for Séamus and the NUJ on workers' rights for journalists. He mentioned collective agreements. Does he want to expand more on that and what can be done to compensate those who have had their work effectively replicated and stolen? Does he think we need to change how we educate journalists? Do we need a radical transformation of journalism courses because of AI? Do they have to change?

My own view is that the AI emperor is naked and we should call it out for that, and that we do not have to redesign our democracy, our education system or how we do business in order to tolerate AI. We need to deal with AI rather than the rest of us meeting the demands placed by the high-tech giants. As the man said, I would not necessarily start from there at all. The reality is that AI is in newsrooms and it is being used there. The education piece is there already.

We need an awareness among media organisations as to how AI is going to be used. In the UK and Ireland we have come across media organisations that are eliminating fact-checking posts and replacing them with AI. In the more detailed submission to the committee we have referenced some work done in this regard, for instance, by Reach in the UK. It is not as if AI is not already there.

In terms of the collective bargaining piece, and the collective representation piece, on my way in I met Deputy Lawlor from Wexford, whom I remember as a National Graphical Association representative who was involved in the negotiations I referred to earlier. Changes in print technology are nothing new. Media organisations and the NUJ have gone through many transformations and we have agreed on the use of technology. Change is the constant. What is different here is that AI was imposed without any consultation whatsoever and it has huge implications in terms of how we do our work and also whether journalists have a job at all.

I was struck earlier by the reference to The Washington Post. It is important. Let us put that in context. The guy who was being hailed as a hero for his war coverage in The Washington Post is now the subject of a GoFundMe collection, which we are supporting, in order to get him home, because he has been sacked by the same people who would present any attempt to control AI as an interference in the right to freedom of expression. We do need to call out some of the hypocrisy around that. I do not know if that answers all of the Senator's questions.

I thank Mr. Dooley. My second question is perhaps for Dr. Culloty or for anyone who would like to answer it. There has been a lot of talk recently about social media bans for young people under 16. It relates to what we are talking about in terms of media literacy. Do the witnesses have any views on that in regard to young people, and how we educate them from primary level, through secondary and all the way up? Do they think bans are the answer? Mr. Dooley has given a view that instead of banning there should perhaps be greater financial penalties on companies. Do the other witnesses have a view on that?

We live in a digital world. I do not think it is right to ban young people from the world that they live in or that we should suddenly expect them to be able to navigate these spaces in the future. Again, it is treating young people themselves and the public as though they are the problem. What we need are safe, reliable, accountable spaces for young people to engage in, be creative in and to have their social lives. Banning is always a kind of knee-jerk instinct when things get very bad. I would rather see more accountability for the people who create spaces that we are letting young people into.

I thank Dr. Culloty. As I have a bit of time left, I will direct some questions to Ms Veldon and Ms Bourke. They spoke a lot about protecting truth and quality in terms of independent journalism and publishers. How do they think the Government can work with them to support that going forward? What role does the Government have to play in that?

There are several things that the Government can do to help us. Primarily, it is around exercising our IP and our entitlement to compensation for the use of our content both on AI platforms and in search and social engines, which has not been acknowledged to date. The EUCD - the copyright directive - is in place. It has been transposed but it is utterly ineffective in helping publishers to assert their ownership of copyright. In the case of AI, we do not even get acknowledgement that the content is being used. At least in the case of search engines, we do, although compensation is slow in coming. There are ways in which Ireland can use its voice, in particular during the Presidency, to assert our copyright. There was a meeting in Europe in May last and it had representation from around 18 states.

Ireland was one of the notable countries that did not speak on the subject while 15 other countries spoke on the topic of copyright saying that it needed to be bolstered so that publishers could assert their entitlement to compensation. We can be helped in that by ensuring that transparency exists from the platforms, that they are brought to the table and that offers are considered and either accepted or rejected, possibly through arbitration if necessary. This is one very tangible thing that the Irish Government can do. The second is in the area of funding, which Mr. Dooley has already mentioned. The funding programmes that have been put in place are, in our view, limited and they are possibly not working in our favour and not working in the right direction. We know the point here is not to compromise our independence. That is more important to us, believe it or not, than it is to the members. We believe there are ways to allocate funding to support innovation, for example in publishing, which are less restrictive than the funding that is going into place at the moment, and possibly more generous.

I thank Ms Veldon.

I thank all of the witnesses for being here. I have questions for each of the three groups but I will see how many I get through in this first question slot. I will start with Dr. Culloty and Ms Chapman, please. I was particularly interested in the opening statement about the AI overhype and how we are being led to believe the level of intelligence of AI by the companies who are building them. That old principle of do not believe everything you read online is essentially being forgotten and in some ways kind of masked or buried by the very powerful marketing of these companies. Ms Chapman has made similar points on the need for a broader understanding by the public of what AI is and its limitations. I think that is being undermined in some ways by these powerful tech companies. From a media literacy and a truth point of view and perspective, how do we push back on this and ensure that what people get from these models is viewed critically? That is really important. I know the education piece is important but should we have stricter rules around disclaimers for AI and how they are marketed? What kind of consumer protection or misleading advertising initiatives could the witnesses envisage?

As I responded earlier to a question by Senator Harmon, the labelling of content often seems like it would be quite intuitively positive but there is very little evidence that it is effective because people have to pay attention to it, they have to understand it and so on. The public just hear about AI now as being AI good or AI bad. There are lots of brilliant uses for AI. There are lots of completely legitimate uses for AI in education and there are lots of great uses for AI in business. I think we mentioned some of those also in the context of news. It is more about this bigger idea that one could just get information from something that is a statistical summary of what is plausible but there is just no factual basis to it. If it happens to be right, that is great but it could just as easily be wrong. It is those things like the summaries and the chatbots being foisted on people that is the problem.

We kind of need a little bit of everything. There is no single solution to any of this at all. It is a little bit of labelling and it is more in the regulation space. We need to be really careful about making any assumption that if we give people media literacy skills or AI literacy skills, they are going to be okay. We probably all have had experiences where we know what we are reading online is maybe not accurate or is not truthful or whatever, but we are still reading it. This is where I talk about a behaviour change as well. Behaviour change takes a long time. Finland was mentioned earlier, I believe by Deputy Geoghegan. Media literacy in Finland has been embedded for decades, not just in the primary and secondary school systems but right from kindergarten up, and even they still struggle as the technology changes. This is not just about skills, it also about knowledge. For example, in the past people might have had a basic understanding of the regulatory regime in terms of radio and television and might assume that if they are looking at something online and it looks and sounds like radio or TV, then it is subject to the same kinds of rules and regulations, whereas it is not.

There is a level of helping people to understand that not all media are equal. This is a really good starting point but we cannot-----

Does Ms Chapman think we should be making the standards equal?

That is really difficult. That is where principle and practice run into difficulties. In principle, from a media literacy point of view it would be a lot easier for people to understand if it was one level regulatory playing field. From a practice point of view, that is really difficult to implement.

Okay. I will move on to the NUJ next, if I may. Apologies if I am going over old ground. I was out at the plinth until now. This applies to NewsBrands Ireland as well but we will see how far we get with it. We know that generative AI and large language models, LLMs, pull from copyrighted material and they do not always give accurate results or attribute credit to authors. From the workers' perspective, how do we as policymakers protect the work of journalists and ensure that these companies are not profiting from theft of their work without their permission, input or control? Maybe we can start with that.

It is very easy to sound like the blacksmith bemoaning the introduction of the motor car and sound like an old fogey. That is not the intention at all, but when the motor car was introduced we did have to introduce rules and regulations and rules of the road, and we had to license and we had to monitor the impact. This is the same way. There has to be a clear and unambiguous enforcement of copyright law. It is a breach of copyright . If the copyright law is inadequate, and we would contend it is, then it needs to be enforced and enforced rigorously. Mr. Cunningham is chair of the Dublin broadcast freelance branch. Our freelance members have been abused for many years. We would say, quite frankly, that the tech giants are free riders on the bus. They are stealing the information of freelancers who are not being rewarded.

The other point I made before the Deputy arrived was that the NUJ has always favoured a tech tax. That would actually have financial consequences and such a fund could actually be enabled to meet some of the costs the Deputy is talking about. The other thing that we do not have in this country is a national training agency for the training of media or journalism. Among all of the many other things that are being thrown into the portfolio of Coimisiún na Meán is that issue. That is a huge gap and there is a need for training of media practitioners in how to use AI. The absence of a formal training system in this country is an issue. The colleges do excellent work but there is also a need for in-service training and training for people who are already employees.

Building on that, and I will come back to NewsBrands in my next slot, in another committee on culture, communications and sport, we heard from a European agency - I do not remember the name of the agency, I will submit it after - which basically had recommendations around a basic income as a potential programme or mechanism for supporting journalism during this really disruptive time that we are seeing. Is that something the NUJ supports?

Absolutely. Journalism is now becoming an unaffordable profession.

Yes. Exactly.

We have had a huge number of people leaving the industry to go into other areas. We have people working as journalists and doing part-time jobs and other work. The idea of some kind of minimum wage or something of that nature is something we would want. Clearly, there are issues around editorial independence and journalistic independence but all of those can be addressed.

I look forward to the Minister, Deputy Patrick O'Donovan, announcing the extension of the basic income for artists scheme today. Will he be extending it to journalists?

I thank the witnesses for all their opening statements. I also welcome our transition year students in the Gallery today. It is great to have a full room. I have an ambitious amount of questions so I will see how I get on. I will start with Ms Chapman and Dr. Culloty, if that is okay. I also sit on the Joint Committee on Children and Equality. We have been talking a lot about empowerment in education and online safety. Obviously there is a crossover with AI. There was a lot of discussion around having one national home for this. The Irish Safer Internet Centre was mentioned. I understand the centre is facilitated by Coimisiún na Meán and that it has more than 250 members. I see that Webwise is under the Irish Safer Internet Centre and that CyberSafeKids also engages with it. The centre also seems to have broader members such as Age Action and radio and journalists. We have so many Departments involved in communication, education, health, justice and we have a Minister of State with responsibility for AI that it seems to me we need a home for education empowerment in online safety, digital literacy and the AI space, which goes under them.

What are the MLI representatives' opinions of it being positioned as this home? I think there was mention of it in the opening statements, but then I realised MLI is a volunteer-based organisation, so does it have any paid staff? What is the number of visitors to its site? How many downloads have their been of its resources, which are fantastic? There is excellent signposting on the website but I am wondering whether enough people are finding it because that seemed to be the problem with the Irish Safer Internet Centre last week.

I thank the Deputy. I will try to answer some of the questions. On the structure of Media Literacy Ireland, it is currently an informal alliance of members - and they can be individuals as well as organisations - who work together on a mainly voluntary basis to promote media literacy. MLI is funded and facilitated by Coimisiún na Meán, but following a recommendation in an Oireachtas joint committee report at the end of 2024, the process is under way for it to be reconstituted so it can receive and distribute funding, which would put it in a better position. The key thing about Media Literacy Ireland is it is truly cross-sector. That is the real strength of it but it also makes it a challenge, because when we are engaged with such a broad range of topics, we can only go into a certain amount of depth. That is why, for example, Webwise and CyberSafeKids are members of MLI but we look to them as members to be the experts in the area of online safety. We see Media Literacy Ireland not so much as a public-facing body but as a business-to-business one, if I can put it that way.

I understand.

Our core purpose is making sure there is co-ordination between different bodies and sectors. That really helps to identify where there are gaps in provision and how we can help to bridge those. A lot of what we do is a bit like a matchmaking service. We identify who has skills in this area, who has needs in this area and see who has funding and resources and bring those together.

What are Ms Chapman's thoughts on the need for one national home for all these entities? It is a bit piecemeal at the moment.

It depends what the Deputy means by "one national home".

I am maybe even thinking of an online Minister. We have a Minister of State for AI already but we have so many Departments that are trying their best. We have this new online health task force and it has recommendations about education. We have the Department of education trying to roll out Webwise but it does not have people on the ground. Then we have that cohort of people under Age Action. I met gardaí yesterday who were stretched in relation to their cybercrime resources. It feels like we need a cross-government approach, but do we need more than we have at the moment?

It probably needs to start at the top with cross-governmental engagement. The Deputy is right that a lot of these topics sit across multiple Departments. We see that with disinformation, online safety and lots of different issues. That cannot really change. AI is another example. These meetings are scheduled in such a way that it is broken into topics and it touches on lots of different areas with different parent Departments, so it is going to be a continuing challenge. The approach MLI has adopted works because we are not engaging with multiple Departments. We are instead working at a grassroots level with the bodies that are at the front line. We are doing the on-the-ground co-ordination, if that makes sense.

Great. In the interests of time I will move ahead to Mr. Dooley and Ms Veldon. There was passionate reference in the opening statement about the platforms that are clearly making money off the back of journalists - because news exists on their sites - and the journalists are not being properly compensated. The platforms are curating through their algorithms and monetising at scale on the back of journalists and then pretending they are merely a notice board facilitating that and not paying the bill at the end. Should we be pushing for legal liability for the platforms as publishers? This came up when we were discussing Grok a couple of weeks ago. Independent freelance journalists who do not have publishers do not seem to have a mechanism under the EU copyright directive to get paid. Maybe both Mr. Dooley and Ms Veldon will give me their thoughts on that.

I have already said they should be treated as publishers. They should never have got away with the notion of deciding for themselves. It is the only sector I know of that could decide to self-identify with no regard for the rights of anyone else. It just does not make sense and it runs a coach and four through the concept of copyright law.

I totally agree. From a copyright perspective, we are not being acknowledged or compensated for the content we are producing that is training LLMs, appearing in AI overviews and so forth. It is critical that changes from a copyright perspective.

On the EU copyright directive we transposed, which you could say we transposed weakly, what are the key next steps for that? Is it a statutory formula for journalists receiving payment?

I think it is the two elements I referred to earlier. The really big issue for publishers is we have no data access, so there is an issue with transparency there from the point of view of the platforms. They are not showing us the data we need to see in order to make a determination about what the quantum would be for payment. The second issue, once you have them at the table, which is not guaranteed - even last week one of the platforms did not appear at an Oireachtas committee - is agreeing with them on some form of arbitration, of which final offer arbitration is probably the most suitable from our point of view.

I thank the witnesses for what has been a very interesting meeting. The theme of the meeting is truth in democracy. There is a saying that truth is the first casualty of war and I hope truth will not be the first casualty of AI. I have a few questions for the NUJ and the witnesses can answer them individually or specifically. What specific transparency and licensing methods would the union recommend?

That goes back to the issue of copyright. Copyright is the clear and unambiguous priority in terms of creators' rights. If your work is being abused, that should stop. Also, all these systems were developed through the theft of our members' work, even training, and we ask for clarity on that. Use of copyright law in the first instance is what we see as the priority.

To continue with that, are the existing EU AI laws sufficient to protect intellectual property and individual journalists? What protections would the NUJ like to see added?

They are clearly inadequate because they are not working. Inadequate law has also been inadequately transposed. The other issue is that the clear and unambiguous identification of tech companies as publishers would deal with a lot of the other problems we are talking about. That they are not recognised as publishers puts them in a class apart and we have to end that fiction.

Would Mr. Dooley like to see media and artists represented on Coimisiún na Meán's board?

Yes. The idea of representation is good, but in the context of this committee I echo the concern that the media industry not being appointed to the original Government-appointed AI committee by a previous Minister of State made absolutely no sense whatsoever. While I recognise this committee has said it is going to be dealing with the employment issue and so on in the next round, the very exclusion of the media from the initial discussion about AI when that was set up, apart from us being able to send a little note, was a clear expression of the departmental failure to recognise that while there are loads of good things about AI, information has not been given due priority. Credit is due to this committee for trying to do this at this late stage.

All right. On media literacy, what do the witnesses think should or can be done to ensure media literacy is embedded across all sections of society?

That is a very big question. Media literacy and people's media literacy needs are very personal and are changing all the time.

We have to look at this as life-long learning. One of the things Media Literacy Ireland has been doing is working closely with the LGMA and the Adult Literacy for Life programmes to try to make sure that adults have access to opportunities to develop media literacy skills in the places they are going and where they are, whether that is a library, a club or whatever else. It is critical that the onus is not on the individual to learn a set of new skills. Human nature being what it is, we will probably not learn them. We have to make it easy for people to develop those skills and there is no one-size-fits-all for that. It is also difficult because traditionally when rolling out a campaign or whatever it is, we look at demographics or target groups. It is difficult to identify clear target groups for media literacy because it is much more about skills than it is about life stages or ages.

What can we do better to protect and improve election period literacy and to protect ourselves from outside interference by big tech companies and countries that are interfering in elections?

During the recent presidential election, the Digital Services Act was in force. The tech companies are meant to have a heightened alertness to the potential threats to our democracy and electoral integrity. EDMO Ireland was part of that process where we were given the opportunity to flag content to platforms if it breached their stated terms or service or things they had broadly committed to, such as labelling AI-generated content. We will publish a report on this in the coming weeks. Generally, we found that it seems quite hit and miss. Some AI-generated content was labelled and some was not. It just falls through, so it is not consistent. That is alarming in an English-speaking country - most of their capacities are most developed for the English language – like Ireland where many of these companies are based, and during an election period when they are supposed to have heightened attention to it. If their capacities appear quite weak and inconsistent in that context, it raises a lot of questions about what they are doing the rest of the time and what they would be doing in other countries.

The other key point on elections concerns the deepfake of Catherine Connolly, for example. The Electoral Commission was monitoring that. It was civil society groups that were monitoring it. We do not really have a proper national service or agency that would be monitoring content consistently. We tend to be focused on elections, but then we also talk about young people a lot. Who is watching and who are the influencers? Most young people are not that interested in the goings-on of the Oireachtas, but there are loads of other people trying to influence them. We are completely leaving that aside because we do not have people whose paid job is to monitor transparently, using rigorous methodologies, and reporting that.

How can we improve time spans in order that something that has gone up on a platform can be immediately taken down? There was the example a couple of years ago of a young girl whose photograph was put up and it was degrading. I think it took four or five years ago to take it down.

In recent years, we had the bizarre scenario of threats against Simon Harris and his family on Instagram and they remained online for many days. Again, that is astounding when you are talking about senior politicians and the company is based in this very city.

I thank the witnesses for their presentations, which sound a loud alarm bell about what is happening and the undermining of the media landscape and the work of journalists. It does not stop there. That has significant impacts in terms of the capacity to have any sort of genuinely democratic society. There is a question about whether that alarm bell is going to be heeded. I appreciate the witnesses' work on it.

This question is mostly for the NUJ and NewsBrands, but any of the witnesses can comment on it. Effectively, big tech has been parasitical on journalism. This is not just an AI issue. It is a social media issue in the sense that content is generated by journalists.

They have to be paid somehow but big tech, through social media, is able to profit from this. It is able to steal the material and put it out there without having to pay any journalists. In a way, AI just supercharges it, but the word "just" understates it. It supercharges the whole thing. You now go to ask the Internet a question via a search engine like Google and unless you choose to opt out, you are kind of force-fed an answer generated by an AI, which has gone to read various articles and come up with something that may or may not be accurate. What is the policy tool to stop this and to ensure we have money so that journalism continues to exist and we have people writing real stories, rather than having the situation, which we are going to get to, whereby AI is generating news stories from other AI-generated news stories in turn and we are down the rabbit hole of AI-generated slop with no relationship to reality because no one is being to go and find out facts about what is happening?

I agree that tech companies are parasites and have got away with being uncontrolled in their activity. Probably part of it is that we are lucky to have them and if they are going to be parasites, we might as well have them in Dublin. I do not buy into that logic at all. I think there has to be an ethical and moral approach to this. My view, which Deputy Murphy mentioned, is that the NUJ has long called for a tech tax. That would be a significant tax, with 6% or 7% the figure mentioned in the UK and Ireland. We are not going to be able to curb these companies so there certainly has to be some legislative basis, but there also has to be a financial fund that compensates.

It is difficult to understate the damage done. We have done surveys on this in the UK and Ireland, and a large number of freelancers have identified that their copy has not just been stolen but in many cases used, abused and misrepresented. We have in copyright law the issue of asserting your moral right, which is the right not to have your copy misrepresented in some way. The treatment of tech giants, as I pointed out to Deputy Gibney, as not being publishers is at the heart of this. They somehow enjoy this self-selected status, and we allow that. That is at the heart of it.

The key point for us is that content cannot be used without consent. That is the biggest issue. Our content is being used without consent or remuneration. We need mechanisms to get transparency, to be able to see data, to see what content is being used and to be compensated for that.

The question of attribution is really important. Hand in hand with compensation goes attribution. That is how you identify where this content comes from, what processes were used to generate it and can I trust it? That is ultimately the question everybody wants to ask. If we allow the platforms to do what they have been doing so far, they will make deals with a small number of selected agencies and news outlets around the world to feed the content and to be their attribution. That will have a severely damaging effect on media diversity and plurality because they will be the only media organisations that continue to exist.

We also need to think about how we get media outlets and the creative sector into the room for the discussions that are going on. We are in the room now but we were not in the room when the AI Advisory Council was developing its paper. We were brought into the room after the paper had been issued and told about its contents. I invite anybody on this committee to read that paper closely to see what it says about asserting the rights of publishers, or any part of the creative sector, to their own content. It seems to us that it is more concerned with the ability of platforms to assert the right to their software IP.

I have a question for Dr. Culloty. Ms Chapman might want to comment as well. I was struck by her strong opening statement, which I agree with. It speaks to how I feel when I see people ask Grok, as if Grok is an oracle or something.

People confuse LLMs for an encyclopaedia. Dr. Culloty spoke about not just scams, low-quality slop and inadequate recognition of AI-generated material but went deeper, noting that a dependency on these unreliable tools undermines our capacity to reason. That is an issue that deeply concerns me. Will she expand on the evidence relating to that?

The evidence comes from researchers at Microsoft. It is a big investor in OpenAI. It develops lots of tools which it is trying to put into schools and workplaces. The researchers found that as soon as people become used to doing something, anything that is convenient and if things are accurate much of the time, most of us think that will be okay. It is not like it is wrong all the time. That makes it even harder for people to resist. Then we have it put there right in front of us. For example, if you open up a PDF document now, Adobe will say "This looks like a long document, let me summarise it for you" and it might be only two pages long. This is being foisted on us. There is growing evidence to suggest it replaces our own thinking. Terms like "co-pilot" are used, which is a complete distortion of what a co-pilot actually is in aviation. It is actually autopilot. It is you not thinking; you are letting something else do the work for you.

It says co-pilot to give the illusion you are still thinking.

In terms of the crossover between media literacy and AI literacy, media literacy involves a huge number of skills. Those skills change all the time. The fundamental building block of media literacy is the ability to interrogate information and make an informed choice based on your ability to evaluate it. Those critical skills apply across AI more importantly than any other bit of the whole literacy tree.

I thank all of the witnesses. I apologise that I was not here to listen to their opening statements but I have read them. I was chairing another committee at the same time. I will first address NewsBrands and the NUJ. There is a particular focus on social media companies in particular being designated as publishers. I do not necessarily disagree with that but how does one square the circle when one looks at journalism as a quite distinct and honourable profession? For example, The Irish Times, Irish Independent or Liffey Champion is the actual publisher of news and content. The journalists who are qualified and registered with the NUJ are the ones actually creating and publishing that content in The Irish Times, Irish Independent or Liffey Champion in comparison with Facebook, for example, where it is a Joe or Mary Soap or a fake account. There is no way we can hold the same standards. I am not saying we should not do so but how can we make the social media companies be designated as publishers when one knows full well the content and quality is not anywhere near the standards a journalise would publish?

A lot of that comes with media literacy. We can differentiate what is journalism and what is factual, investigative content compared with content anybody else could produce anywhere, from a blog to a post. A lot of people do not differentiate. They get their newsfeeds in so many different ways now. Society has changed. The way people consume content is very different. Rather than classifying as a publisher, it is more about having legislation and rules around what we deliver. We work with the Press Council of Ireland and Office of the Press Ombudsman, which we fall under.

It is about having the same type of legislation to ensure the social media companies are held accountable in the same way publishers are.

As Louis Armstrong said about jazz, if you had to ask what it was, you would never understand. The truth is a bit like that as well. In terms of publishing, the distinction was made earlier between news and information. Journalists adhere to the NUJ code of conduct and media organisations adhere to codes of practice. I wish it was the case that journalists had to be a member of the NUJ but journalism is not a registered profession like doctors and teachers. Therefore, it is quite hard to define journalism. I am not advocating that people have to register or must have gone on one course because that would interfere with plurality and diversity. However costly it may be, I favour the line of "publisher" because it makes it clear that if you steal other people's work, use it and put it on a platform, that is a publication. They may say it is too costly and involves too much effort. I challenge the ethical basis for saying that because we will earn less money, we are not going to put in checks and balances. That is where we are coming from.

I think there is probably a middle ground. I am not sure if "publisher" is the right terminology but there need to be more guardrails. A tech tax was mentioned. It could be along those lines and more of a fining mechanism for disinformation or whatever it may be. There are publishers here as well. Journalists are making money off social media sites as well. People share content and articles on Facebook, Instagram or wherever. They benefit from it as well. That is a difficult one for journalists. Would their content get out as much if the tech or social media companies were regulated more? There are publishers in Ireland that use clickbait on social media. Would Mr. Dooley agree? This is the reality. We are criticising social media companies, rightfully so-----

From a journalist's perspective - publishers can speak for themselves - vanity does not butter parsnips. Getting your name on Twitter is great but it will not pay your rent. The notion that somehow you should be lucky to get it because you end up on X is a bit of a mirage. The other point to balance against that, speaking from a journalist's point of view, is that the failure of social media platforms to protect media workers, for instance, on X, is also a major issue. There are particular issues around misogynistic abuse.

A lot of the content is being taken and misused from journalists and online publications in particular is because of crawlers. Crawlers have been going for at least 15 or 20 years. Factiva is used in the Houses of the Oireachtas. Things are behind a paywall and you pay for it. Do the witnesses have any suggestions around legislating for crawlers and how they are used or misused? We will probably need to look at that. It gets to the crux of the issue of information and articles being misused and, in a way, stolen and presented in a different format. Have the organisations looked at that?

Previous licensing arrangements generally adhered to the copyright laws that existed. They did not present a problem for us because we had the licensed use of our copyrighted content. On the question of our content being used on social media and in search, we do not have a choice in that. We would not exist now if we were not able to get our content out through those mechanisms. For all that, we are saying we would welcome further regulation of those, as publishers if needs be.

At the moment, they have the luxury of being able to carry the content without any of the responsibility that goes with that, which publishers take on every time they publish a story.

I thank Ms Veldon very much. It now falls to me to thank the witnesses for their contributions. I have found it fascinating. I am going to be deliberately provocative. I am not allowed to ask, but we have a number of transition year students who are here with us. We often go into schools. I am a news junkie. I wake up and read The Irish Times and the Irish Independent. I look at RTÉ news. I absorb it. However, when we talk to those, say, under the age of 40, for the most part, the traditional media is not where they are finding their news sources. They are far more likely to have a social media influencer than an opinion page writer forming their views. In many cases, increasingly, some of that content is AI enabled.

My question is, to be provocative, whether the battle is lost. What can we do to engage with a younger generation? We are going to make a series of recommendations around this committee and so on. However, I say respectfully to my colleagues that our thinking is probably formed in a very different way from that of a newer generation that is coming up. We need to engage with that generation a lot more. What are the witnesses' views?

There is not a newsroom in the world that is not grappling with the question of how to engage young audiences. There is a struggle going on there. We need to make no mistake about that. As for whether the battle is lost, I do not think so. We see lots of inventive and innovative activity where young audiences are being engaged, particularly through the medium of audio, which they seem to respond particularly well to. We have not given up the battle yet, but we do need to accept that AI will change the news ecosystem fundamentally, and we need to be part of that redesign.

The Cathaoirleach is absolutely right in that there is a massive disconnect between people talking about politics, AI literacy, news literacy and all these things and the actual lived reality of most people. However, I also think that was always true. I do not particularly remember spending lots of my time following Oireachtas committees when I was 20 years old or so, so we should not overstate it either.

Dr. Culloty does not know what she was missing out on.

The important thing is that in Ireland, we are actually quite exposed by virtue of being a small English-speaking country. An awful lot of those influencers who are really prominent in online spaces are coming from the US and less so the UK, but very much the US. Health and wellness and gym culture and beauty culture are massive areas that we are often blind to. Here, however, there are basic regulatory principles like consumer protection and platforms incentivising certain types of content. Those are regulatory principles that we can look at. In terms of young people, we have been doing some work. Young people getting involved and doing things in their local communities, whether that is through sports or arts or whatever, is how they find out inadvertently about the news and get engaged in things. That is another way to think about it instead of a purely technology-based thing.

I agree with the Cathaoirleach that this is a challenge for everyone and a challenge for democracy. The real challenge is that if we accept that then we mean that the views of young people are being shaped by algorithms directing preconceptions into an echo chamber. I agree that the misogynistic and racist material they are driven into are all problems. However, young people also have an ability, as new generations of users, to differentiate between truth and lies. The global interest of many young people in what is happening in America and Gaza and all of that presents a challenge to media organisations not to give up the ghost and to invest in editorial resources and try to fight that battle. Perhaps it is a question of failing better, but we must continue.

Would Mr. Dooley accept an AI chatbot journalist as an NUJ member?

Absolutely not, no. I have enough problems with the members I have already.

I will maybe just add something from a youth perspective. We worked with a research company called Colourtext to really look at the youth because something we are focusing on in publishing houses is how we engage with the youth.

One thing is that we sometimes underestimate what they think. About 76% of respondents said they were concerned about fake news and what they were reading. They were seeing things come up on their social feeds. Therefore, they are turning to traditional news to verify it because they have been caught out. They have seen news and then they are finding out it is not right. We should not underestimate that the youth are looking to verify.

I would like to flip it around a little. If we look at a lot of studies, the problem is not that young people are susceptible to misinformation or AI disinformation. The problem is actually old grey-haired people like me who are not digital natives, for want of a better phrase, and are way more susceptible and much more liable to spend time falling down rabbit holes on Facebook. We need to think about what we are doing with young people and how we are reaching them, but we also need to think about how we are reaching older audiences, who are spending more time online. They have grown up all their lives knowing news as reliable when what is presented as news is not reliable. That is a much bigger challenge for them. They have been trained for a lifetime to accept it.

Mr. Cunningham kind of stole my thunder there. I am actually much less concerned about school-aged people. There are great resources in schools already. A different way of thinking is developing. My concern is the 20 to 40 age group because they are probably more engaged, more active and more vocal online. Again, it is about reaching those kinds of people because we do not have a very convenient way of reaching them.

I want to throw one quick question to Ms Bourke because of her experience of the Australian media. This committee will be engaging with and looking at Australia's response to AI and the relationship with tech companies as well. Does Ms Bourke think there is anything we can learn here in Ireland, particularly in the regulatory space? I am not necessarily talking about the current discussions, but the relationship between the traditional media and the problems of scraping that she and Ms Veldon talked about. Obviously, Australia took a fairly strong stance. I am thinking about any policy recommendations from her here.

The Australian Government really got behind the concept of copyright and backed the media organisations from a democracy point of view regarding content to support them in ensuring they could get to the table. It also put legislation in place to ensure that where content was being produced, which costs a lot of money to do, the publisher is being remunerated for that content. The intervention was very important.

We can go for another round of four minutes per member, so I will choose members who are still here.

I thank the Chair. I am sorry if I ask some witnesses the same question, or the same witnesses. It is not that I am not interested in other witnesses. It is just that we have limited time.

Both Ms Bourke and Ms Veldon mentioned how the EU Presidency could be a great opportunity for Ireland to plant its imprint. Then, however, Ms Veldon was very critical of some meeting that took place where 18 member states were present, and Ireland said nothing. Could she elaborate on how we are going to square the circle between how bad she thinks the Irish State is doing at advocating on the copyright side versus what is forthcoming in terms of the EU Presidency?

In fairness, that is probably not a typical experience, but what we as publishers see is that there is lip service to the notion of copyright. There is an acceptance at Government level that we do own our own content and that we should be remunerated for it, but we do not see any further action on that. That says to us that the European Union copyright directive, EUCD, needs to be looked at to see if it is powerful enough. Its transposition needs to be looked at. We need to look at other elements now that are coming in, such as the AI Act and the Digital Services Act. All of these are going to impact on our ability to fight for our rights, essentially, and the Presidency offers us a very good opportunity to do that.

I will come to Ms Bourke, but Ms Veldon mentioned some arbitration that is ongoing and indicated she hoped the arbitration might result in better outcomes, or that is the process.

What can Ms Veldon say about that?

I am sorry if it was not clear but that is really the process. We are looking for two things from the platforms. One is transparency in sharing the data they own that is about our content so that we can at least have a discussion about how we might be remunerated for that content. The second thing is about how we come to a decision as to how much is to be paid to publishers of that content. One of the proposals on the table at the moment is final offer arbitration. This involves both sides coming up with a figure based on the data, which would now have been shared, and that is then subject to the decision of an umpire or referee. That is called final offer arbitration.

Excuse my ignorance but who is proposing this? Is the European Commission proposing it?

It is being considered in Europe. News Media Europe, which we are part of, is lobbying at a European level for final offer arbitration to be accepted.

Would that come in through the transposition of the EU copyright directive or is it a separate proposal?

It would be a change to the regulations which would then permeate down to the individual member states.

Ms Bourke mentioned the AI Act and how Ireland could become a leader. How does she view the Act in its current form and how it is going to be applied as it relates to her industry, leaving aside the copyright directive? What is in it for her?

A few main points in the Act stand out for us. For us, it is also about the implementation and safeguarding creators and rights holders. The requirement for AI companies to publish the details all comes back to the copyright. The largest component of implementing and enforcing the EU AI Act here in Ireland is transparency and getting the data, which Ms Veldon mentioned. It is about getting to know what is actually being used. We need to have access to data on anything that is being used to train the LLMs. We need visibility of that. This should allow publishers to be compensated. It is about a mechanism to get data and to have transparency.

I am not asking the witnesses to be Nostradamus, but if we came back in three years' time, would the publishers have been compensated and would they have been given access? What is the direction of travel in Europe or even in the United States on the crucial points the witnesses are raising?

The experience is varied. In general, countries that are taking up the cause of publishers or recognising the importance of the continued existence of the media industry are having better success. They are joining cases. It is through litigation that a lot of this will ultimately be sorted out. There are things like the Google adtech case, which is being appealed. That depends on getting support from lobbying groups right across Europe and from individual member states, where that is possible. It is a varied experience. I would like to be a little bit more optimistic than I actually am but there are a lot of cop-outs for AI platforms at the moment. The experience with search and social and the fact that copyright was not comprehensively dealt with at that time mean we are now completely on the back foot when it comes to something as complex as AI, which not only takes content but distorts it beyond recognition.

I know I said I would come to Ms Veldon but she has actually covered what I was going to ask her so I will raise a couple of new items. I have often quoted a book called Rewiring Democracy by Bruce Schneier, which came out recently. There is a section in it on informing the public. This is where he deals with AI's impact on traditional media. The book is quite challenging for me because it quite positive as well as being sceptical and concerned about the dangers. Mr. Schneier does talk about some of the ways in which AI can help journalists and politicians to tell their story. However, he also talks about some of the more negative impacts.

There are things like NotebookLM, which is a product that can replace podcasts with fully automated voices. The author also talks about the technique of flooding the zone and the personalisation AI will produce so that people's news feeds will be not just customised but specifically personalised. Will the witnesses speak about those very cutting-edge ways in which AI is interacting with the delivery of news? This is mostly directed at Dr. Culloty and Ms Chapman but I would like to hear from whoever wants to answer. What is coming around the corner? Of these things, what is really of concern to the witnesses?

In a way, a lot of those things are already there. There is already a capacity to highly personalise content, especially through algorithmic feeds. The thing that most concerns me when we think about democracy is that the whole AI issue is now heavily tied up in geopolitics with countries whose leaders are not committed democrats. The people who run these companies are not even committed to the human race, never mind democracy. It is a bizarre scenario to be in. It is really hard to get a handle on that. It is just so different from where we have been. With algorithmic feeds, we can already see the way in which certain content and ideas are pushed. The roll-back against democracy is my biggest concern.

Last week, in another committee, I asked representatives from TikTok whether they thought there was interference from the Trump Administration in the recent roll-out of the new TikTok model for the US. Was that the case? The representatives denied it flatly.

It would appear so. The capacity to demote certain content and ideas has been a problem for quite a long time. We also know that, in the case of Gaza, some platforms were systematically hiding or demoting content from Palestinians.

And Epstein is currently the-----

This has happened in respect of loads of topics.

One of the things that concerns me about all of this is that media plurality is one of the pillars of democracy. We have lots of checks and balances and systems in place to make sure media plurality is maintained. When we look at the situation the Deputy is describing, that is gone. That is a fundamental issue. Even if somebody has those key critical thinking skills, if they are not getting the content to engage with-----

They are not being challenged.

-----and are not getting alternative perspectives - I am referring to alternative perspectives rather than alternative facts - then they are not being challenged. There is potentially a job of work to be done to make sure that there are some checks and balances in place to make sure people have easy access to alternative perspectives.

Does anyone else want to answer on that particular point?

On the question of moderation, the committee tried to get to the heart of this last week, although without success or real acknowledgment. Human content moderators on the platforms are being replaced with AI. That is going to be a bad thing. Whether it is acknowledged or not, it is happening.

It is. I worked in the tech sector from 2006 to 2013 and, during that time, the handling of certain sales and customer services processes were always moving towards automation. They are always going to be because it is about profit and the bottom line. Of course, that is going to continue with these processes as well.

This morning's theme is truth and democracy. It is so important for the young people behind us to focus on their critical thinking and reasoning skills. That is very important. I have a general question. Do the witnesses think AI is capable of replacing artists in producing work, music, poetry, painting and, as has been mentioned, photography?

There is every danger that there will be an attempt to displace visual artists, photography and voices. However, it will not replace them because it lacks the creativity and imagination.

That is the case at the moment.

Whatever else it will be, it will not be the product of human endeavour. It is designed to replace human endeavour. In that sense, I have to hope we will ensure that does not happen.

What we get from AI is not reality or a human creation. This is why we mentioned the values of the European Trade Union Confederation. Whatever happens in AI, it must be human led. There must be a human dimension to it. I am optimistic that if we go down this road then, for all of our faults, collectively across the globe we will not allow AI to replace Beethoven, Heaney, Brenda Fricker or whoever.

What newsroom-level standards should be mandatory when deploying AI in reporting?

My view is that AI is not an appropriate mechanism for reporting. If it is to be used, it should be to provide statistics that are then overseen and analysed. It cannot be used to report. As far as I am concerned, it would not report in an independent journalist sense.

What immediate reforms would create a fair bargaining environment with big tech, particularly on copyright and content licensing?

That might be something for NewsBrands in terms of licensing.

From our perspective, with regard to the AI Act and the copyright framework, it is about making sure they are fundamentally fit for purpose. AI is changing the world every day and the legislation constantly needs to be reviewed. From the perspective of a free press and thriving democracy, we would like to be involved in policy development. It is critical that we are engaged in the process.

We will do a specific module on AI and the creative industries. I imagine that quite a number of these issues will also feature then. One thing with regard to the committee in a cross-party way is that we have been very committed to the idea that free, fair and balanced media are an essential part of our democracy. Perhaps we take it for granted in Ireland. In the week that Jimmy Lai has been jailed for 20 years in Hong Kong, it highlights the importance of what we have here. With all the flaws, blemishes and problems, it is something we should celebrate.

The committee makes recommendations and looks at where the Government is moving. I want to speak about some of the skills. The Minister, Deputy Lawless, has been speaking about the national skills observatory, which will incorporate the AI observatory that the committee has been speaking about. There is a plan to roll out a series of education programmes which will be open to the general public. Some of them will be micro-credential, ranging from explaining what AI is to how people can embed AI in their business or how they can use AI in creating content, by which I mean doing it responsibly. There are responsible ways in which journalists and others can use it, as Mr. Dooley has outlined. If the witnesses were to make recommendations on these skills and where we are moving in terms of equipping Irish people to be AI literate, what are the recommendations they would like to see?

In my university, DCU, a lot of work has been done on transversal skills and the idea that no matter what people study or what area they might like to go into, there are core things that should be developed, such as critical thinking and ethical literacy. Critical thinking is the foundational component. People will pick up AI "how to" skills. If they are interested in something, they will pick up the "how to" skills. It is the ability to question things and think critically that will be key, no matter what sector something might be in.

I would love to see a national longitudinal survey on media literacy skills, digital literacy skills or whatever we want to call it. Until we know where the skills needs are, it will be very difficult for us to create interventions to meet those needs.

Does anyone else wish to speak?

I agree that critical thinking is fundamental. It is important. There is nothing new in this. I said earlier that the emperor is naked. In the Hans Christian Andersen story, the first journalist to call truth to power said the emperor was naked. That is the first example of critical thinking-----

It was a child.

-----and a departure from groupthink. That fundamental departure from groupthink is what we need. The groupthink at the moment is that we are afraid to sound too fuddy-duddy if we challenge AI. No one is saying that there is not potential with AI, but we should not be afraid to be critical of its power and those who refuse to recognise that power.

I credit you with great spin there. It was a child, rather than a journalist, who called the emperor naked. I will allow you that.

We have very young members.

We talk a lot about media and AI literacy and young people, in particular, not realising that they are being manipulated. What is probably scarier in many ways is that all of us are using these outlets. We are all being bombarded by them and are subject to them. We are all being manipulated. The skill set of learning how to manage one's time on one's phone and involvement with social media on any platform is important.

On skills and developing interventions around them, it is important to be mindful that just because we create opportunities for people to develop skills, that does not necessarily mean they will take up those opportunities. When we are looking at creating those interventions, that is where the research and insight comes in so that we are able to create interventions with which people will engage. The Senator is correct. This is about behaviour change for all of us.

I thank the witnesses for their useful contributions. It will feed into the work we are carrying out. Our committee always welcomes additional contributions, in case people want to email ai@oireachtas.ie with any research. That concludes today's meeting.