Wednesday, 22 November 2023
Artificial Intelligence and Emerging Technologies: Motion [Private Members]
“That Seanad Éireann:
notes:- the enormous challenges and opportunities that Artificial Intelligence(AI) and emerging technologies present for all sectors of society;recognises the positive environment for AI research and enterprise development in Ireland, underpinned by:
- that AI will be embedded in almost every aspect of life;
- that AI could lead to significant positive improvements in delivery of services such as healthcare or in transport;
- the particular impact of the use of AI on elections and on the democratic process and how it can contribute to the spread of misinformation and disinformation;
- the development of the AI Act as part of a suite of digital legislation and regulation by the European Union; and
- recent global discussions on AI regulation, including the UK Government’s AI Safety Summit;- the National AI Strategy: AI –Here for Good;welcomes:
- the support of State agencies for indigenous and multinational AI and technology companies; and - the positive work of ADAPT, the world- leading Science Foundation Ireland funded AI research centre involving Ireland’s universities;- the establishment by the Government of an AI Advisory Council and the appointment of an AI Ambassador;further notes:
- the plan by the Government to establish an AI cluster to support enterprise development in this sector; and
- the development of tax incentives for angel investors in new technology start-ups;- the need for public engagement and greater understanding of Artificial Intelligence, emerging technologies and their impact;and calls for:
- the potentially destabilising impact of the misuse of AI and deepfakes in elections and referendums;- the establishment of an Oireachtas Special Select Committee on AI and Emerging Technologies during this Oireachtas term;
- the development in the next Oireachtas of a Committee of the Future, similar to the Committee of the Future developed in Finland, that will examine and report on major national challenges and opportunities, including through technological change;
- greater emphasis on understanding of AI and other technologies within our education system;
- the development of a framework for how the Government can collaborate better with academia and industry to guide AI research and the ethical application of the technology;
- a training programme in the Oireachtas on the impact of technology;
- a national awareness campaign on AI and technological change and its impact;
- a strategy to be formulated by the Electoral Commission to tackle the misuse of AI in elections and political campaigning; and
- a code to be developed for all political parties and candidates around the appropriate use of AI in elections and political campaigning.”
I am grateful to the Minister of State for coming to the House this evening for this debate on artificial intelligence and emerging technologies. We are in a period of tremendous change. Artificial intelligence and other technologies are changing and will change our lives in a very dramatic fashion, now and into the future. There will hardly be a sector in Irish society or globally that will not be impacted by this. Therefore it is critical that all our citizens are prepared for those changes and that everyone is poised to benefit. This motion recognises that the Government has taken the lead in several areas with regard to artificial intelligence. This includes the development of the AI strategy. It is two now years old and given how quickly AI and new technology is changing there is probably a need for that strategy to be more dynamic and constantly updated. The Government has appointed an AI ambassador and there is a process to appoint an AI advisory council. I was very glad when the Minister of State informed us there had been a lot of applications to sit on that council. He might update the House on that.
It was particularly welcome that the Minister of State personally attended at Bletchley Park in the UK for the AI safety summit. I also welcome his support for my calls for the idea of Ireland hosting the EU’s agency on artificial intelligence.
While much of the concentration around artificial intelligence and machine learning will be about potential harms, it is also critically important that we recognise the many opportunities for good that can result. Healthcare is one of the particular areas where it could be transformative whether in diagnosis, treatment and clinical decision making or, indeed, in drugs research. It is essential that there is an ethical basis for the use of AI. That should be no different to the use of any new technology. It is also important when we talk about healthcare and AI that there is always a right to a second opinion and that is maintained. Algorithms already out-perform radiologists identifying malignant tumours and cancers and large data models are being used to assist in clinical trials. That is just one area where it is positive. In every single area, from education to agriculture, it will be transformative.
Therefore it is important that with this new technology the necessary safeguards are in place to put the citizen first. The AI Act being discussed at European level has the potential to be the most significant piece of legislation enacted by the European Union this decade. Given the capacity of AI to cause harm as well as good, it is essential that the EU does reach agreement and that the risk-based approach, with a view to minimising risks to individuals and humanities would inform the final legislation. It is critical that we would not assume that technology companies will conform with any regulation. Indeed, if our experience with social media has taught us anything, self regulation by the technology companies does not work.
Before the deployment of any new AI-enbabled technology takes place, we must ensure that risk management takes place and that this must be independently assured. We need to remember that with AI most flaws cannot be removed once a system has been trained. It is critical, therefore, that we enforce AI product safety in the same way we would expect any other product that would be brought to market.
As part of the Spanish presidency there have been efforts to come to a final agreement around the AI Act. The principles of a risk-based approach and the precautionary principles as well as the principles that underline our own national strategy about the safety of citizens and putting citizens first must also inform the AI Act.
I was particularly glad to see the Minister of State was at Bletchley Park because democracies have a particular responsibility to ensure that human rights and ethics inform the roll-out of any AI technologies. We have to protect our democratic institutions and elections from the misuse of AI. The spread of misinformation and disinformation in elections is nothing new. It has gone on back through the ages but the difference, as we all know, is that AI will now allow it to be done at a scale and speed that has the potential to destabilise the political environment in which we all operate. We have seen the deployment of deep fakes and synthetic media in recent elections including in Slovakia and in the presidential election in Argentina. Some of these deep fakes appear to be relatively simple constructs but as with all technologies, they will improve over time and will become far more believable.I have absolutely no doubt that in electoral contexts in this country over the next two years, we will see the deployment of deep fakes, the misuse or attempted misuse of AI, and the spread of misinformation and disinformation, which could potentially damage our political process. I have no doubt that at a crucial point in an election campaign here, there will be an effort to deploy a deep fake of one of our political leaders or of somebody who could influence an election and that could have profound implications. People in this House can only imagine what would happen if the Tánaiste, the Taoiseach, Deputy McDonald, Deputy Bacik, or any of our political leaders were to appear through a deep fake, three or four days before an election, to be saying something that was controversial or deeply damaging to their election prospects. The difficulty is that this deep fake would be widely shared on social media and by the time a denial was issued, the damage could potentially have been done. The discussion in the political context would be around the deep fake. There would be plenty of people who would deny, even if it was identified as a deep fake, that it was a deep fake in the first place. We need only look at our experience during the pandemic with the spread of misinformation and disinformation and think about that supercharged in an election context. Obviously, that is not just going to have an impact on elections in this country; this is something we will see happening globally. Therefore, it is critical that the Electoral Commission take this matter quite seriously. There is a responsibility on political parties to sign up to a pledge that they will not misuse AI for the purpose of deep fakes or for misinformation or disinformation. There is always going to be a danger of campaigns that will do it. There will always be a danger of state actors, particularly from non-democratic countries, who will use it to try to destabilise political democracies and we have to be on our guard against it. I would certainly hope the Electoral Commission will regard this as a top priority.
We very clearly need action from the technology companies themselves. We have to ensure that when they are designing any product, they adopt the original Google motto of "do no evil". We have to think about ensuring when those products are being designed that product and citizen safety is built in right at the very start. However, as I said before, we cannot allow self-regulation to take place. I totally accept that we need to allow regulation and legislation that does not hinder innovation. We must have regulatory sandboxes that allow new experiments and new innovations to take place, but at all times we have to ensure the safety of our citizens comes first. Striking that balance is critical. As I said before, it is no different from having regard to product safety when rolling out any new product.
Our education systems will need to be adapted so that they can make effective use of new technologies and so that those students going through our school systems understand algorithms, how deep fakes are made and their potential impact, and the potential to be able to solve global challenges.
The Minister needs to lead on this debate, which is around a process and around improved public awareness of the impact of AI. I have previously called for an Oireachtas committee or a special Seanad committee on AI and emerging technologies, with public hearings which will broaden the public debate. In the next Oireachtas, we should consider establishing a committee of the future along the lines of that established in Finland, which looks at how big-picture issues will impact on Irish society. I thank the Minister of State for his leadership on this issue. I hope we can contribute further to the debate on something that will transform all our lives.
I certainly welcome the opportunity to second this motion in the House. I thank the Minister of State for being here and for his sterling work in the area in appointing Dr. Patricia Scanlon as AI ambassador, and in other initiatives he has taken such as the establishment of the enterprise digital advisory forum, the AI innovation hub, and of course joining the Global Partnership on Artificial Intelligence. This is all very valuable and important work. Artificial intelligence raises very important and urgent questions for all of us as legislators and for wider society. Of course, artificial intelligence is already with us. It is not a concept we are looking at in the future. It is changing the information we receive, the choices we make and the ways in which our societies function. There is no doubt that in the coming years, AI will play an even greater role in how governments and public institutions operate, and in how citizens interact and participate in the democratic process. AI is an opportunity to improve the democratic process in our societies. For example, it can help citizens to gain a better understanding of politics and engage more easily in democratic debate. Likewise, it can aid all of us as politicians to get even closer to citizens, understand where they are at, and hopefully represent them more effectively. Such an alignment between citizens and politicians could change the face of electoral campaigns and considerably improve the policy-making process by making it more accurate and efficient. However, there are very significant risks associated with AI from a democratic perspective. AI has the potential, as outlined by my colleague, Senator Byrne, to operate as a powerful tool for disinformation and misinformation and we all have to be acutely aware of the hugely negative and divisive impact that these can have on the body politic. Notwithstanding the risks, and having spoken to experts in the area, AI could prove useful to democracies if proper safeguards were applied. I understand specific tools can be deployed to detect the use of AI-generated content and techniques such as watermarking can be used to clearly indicate that contents have been generated by AI. AI systems are also increasingly being used as part of risk management strategies such as in electoral content moderation. Social media platforms are also increasingly using AI systems to detect suspicious patterns in content before elections. All of this is important and in this way AI systems can make a valuable contribution to the detection of fake news and to support citizens in reaching informed, factual-based opinions.
In my role as leader of the Irish delegation to the Council of Europe, I have engaged with this issue on an international level. This is obviously far more than just an Irish issue. It is an issue that requires a collaborative regulatory approach right across the globe. The Council of Europe is mobilising to develop norms that are adapted to the challenges encountered, including a framework convention on artificial intelligence. Most Council of Europe committees, intergovernmental bodies and specialised bodies, as well as its monitoring structures, are considering the impact of AI on their fields of activity and it is important as a State and as a Government that we engage fully in that process. This is one of the three priorities I recommended as rapporteur for the fourth Council of Europe summit last May and was accepted and adopted by all 47 member states.
As party spokesperson on education, I am especially intrigued by AI's potential benefit to our educational provision. AI has a variety of algorithmic applications in education such as personalised learning systems to promote student learning; automated assessment systems to support teachers in evaluating what students know; and facial recognition systems to provide insights about learner behaviours. Besides those platforms, algorithmic systems are prominent in education through different social media outlets such as social network sites, microblogging systems and mobile applications. These AI systems can increase the capacity of educational systems and support the social and cognitive development of students and teachers. Therefore, there is certainly huge scope for AI to improve our education systems, not just in Ireland but around the world, however there are significant risks associated with AI in education.The biggest risks associated with integrating these algorithms in educational contexts are perpetuating existing systemic bias and discrimination; perpetuating unfairness, affecting students mostly from disadvantaged and marginalised groups; and amplifying racism, sexism, xenophobia and other forms of injustice and inequity. Academics point out that algorithms do not occur in a vacuum but, rather, shape and are shaped by ever-evolving cultural, social, institutional and political forces and structures. There is a concern that over-reliance could have negative impacts.
It is welcome that we are having this debate. I know it is going to be very difficult to regulate in this area. As the Minister of State accepts, it is really important to have a strong regulatory framework at national and international levels that will harness the potential of AI but also absolutely protect society from the potential pitfalls.
The Minster of State is welcome to the House. I thank my Fianna Fáil colleagues for introducing this motion, although they seem slightly obsessed with AI and its use for elections and the like. Let us be honest about it: I can think of no situation where we will have an artificially intelligent Minister of State sitting in front of us answering questions in the House. As such, we as politicians will not be easily replaced by AI at any time in the foreseeable future.
To leave aside the jokes and get down to the serious issues, AI has the capacity to rapidly evolve cybersecurity risks within our country. Not only is our country suffering from these, so too are countries right across Europe. The level of cyber awareness that exists within populations is desperately and deeply worrying, as is the replication capacity of AI. We saw just today in The Irish Timesthat a teenager was responsible for cyberattacks in this country. That is just frightening.
Many will know I have put together an expert group that meets in this House from time to time. Recently, we established a common statement or vision for cybersecurity in Ireland, and that vision encompasses AI. Many Departments are already using AI to some degree. I have some concern about the rush to regulate. While we will try to regulate, we will be one step behind the criminals using AI whether we like it or not. Therefore, it should be exploited for all of its worth and we should learn what criminals are likely to do with it once they get their hands on it. Believe me, they are working on it at the moment.
There is a misconception that criminals involved in cyberattacks and the use of AI are super-gurus when it comes to IT. They are not; they are people who get lucky. They have a specific algorithm they are able to work on, or they buy the cyberattack software on the dark Web.
At a recent conference, we were talking about customer service and the obligation of those supplying goods and services to provide quality goods. If you buy cyberattack software on the dark Web and it does not work, there is nowhere to go with it. However, this is something we need to be extremely serious about.
In our own Parliament, there is a question over the CCTV systems and the use of Chinese hardware. Has this House been observed from outside? Who knows? This is something we have to be aware of. In the United States, TikTok is now banned. We have got to decide in this country whether we want to go the same way.
Very recently we had the director of the National Cyber Security Centre before our committee. One of the problems hindering the centre right now is the lack of legislation. There are all sorts of things to be considered when talking about legislation that will govern not only the use of AI but also cyber defence in this country. At the end of the day, it is very difficult. Senators Byrne and O'Loughlin are correct that systems are moving forward at a pace and that it is extremely difficult for a government ever to get ahead of them. Back in the days when I was teaching information technology, one of the issues we had was that hardware was finding it hard to keep up with the development of software. We are going to run into that as AI takes over. The types of machines needed to run AI are going to be pretty enormous.
Senator Byrne referred to the large companies that have a vested interest in AI and everything to do with it. At a conference I attended, Google spoke about its vision to be the educator of the world – because you could access anything you wanted on Google – until somebody questioned whether Google was censored in China. Of course it is.Therefore, it is not free and open when it comes to these things. The truth of the matter is that these large companies have a vested interest. We are now negotiating with the likes of Meta as if it were a nation rather than just a company. This is deeply worrying. The owner of X, Elon Musk, is being called into parliaments to outline where he sees AI going. This is deeply concerning to me. We should be using our own experts, not relying on the expertise of major multinationals with a vested interest in ensuring governments move their way.
I am delighted today, after nine years of calling for the establishment of a national intelligence and security group, that the Tánaiste has announced we are about to put in national intelligence. We are the only Parliament in Europe that does not have an intelligence sub-committee to examine and oversee intelligence in the country. The move is really good.
I am delighted to see that, under the leadership of the Minister of State to a certain degree, the National Cyber Security Centre is about to move into its new premises. It is increasing its staff number. Funding is the main issue for it. We have got to see it funded as much as possible. I am running out of time.
The bottom line where the National Cyber Security Centre is concerned is that we cannot put enough money into it.
My final plea to the Minister of State is that the Government fund industry with respect to cyber awareness. Our nation should be the most cyber-aware country in the world given the amount of FDI in it.
I welcome to the Gallery guests of the Minister of State at the Department of Finance, Deputy Carroll MacNeill. They are most welcome to Seanad Éireann. I thank them for coming here. I hope they enjoy the debate and their stay in Leinster House.
I welcome the Minister of State. It is a particular pleasure to do that, having served with him at the Council of Europe in the past. I salute how avant-garde he is and how much he is embracing this whole area, leading out on it and ensuring we are at the front in the area of regulating the use of AI both in the positive and controlling senses.
I welcome our joining of the global partnership and the investment in the National Cyber Security Centre. I salute the Minister of State for all this. It is a great pleasure to hear from him. I look forward to his response, which will be very interesting.
I am happy to welcome and endorse the Fianna Fáil motion. I congratulate the party on conceiving a motion that is very much of the moment and the future. We support it.
The first thing that merits saying about AI is that it represents an exciting new vista.Looking at it from a positive point of view, it is generally understood that it will have phenomenally good effects in the area of health, that much more intricate surgeries will be able to be done with extraordinary precision and perhaps much rapidity, and that it will greatly enhance our health service. That is understood, is a given and is widely accepted. It has wide implications in the areas of agriculture, education and right across life. There is a very positive side to AI and to its development and it is effectively for good.
Due to its sheer power, its newness and the risk of it being controlled by unethical actors or by people of malevolent intent, it is important that it is regulated and that there is control. In that sense I also welcome the EU AI Act and I hope that it evolves in a satisfactory way.
I also agree with Senator O'Loughlin in that the Council of Europe is out there suggesting controls and regulation. The regulation and control are important. I believe it has to be controlled by parliaments, by the EU, and bodies like the Council of Europe, which have to set standards and criteria and make demands of member states in that area. Regulation and control will be important.
Senator Craughwell cautioned against an over-emphasis on the threat to democracy but I do not know if I fully agree with him on that because I believe it is important to put an emphasis there. It is important that we would be conscious of the risk of what misinformation and disinformation could do, and what could be done through deep fake. There is a risk there. We have seen - before the full unveiling of, development and full access to AI - the way in which there was a manipulation of the UK electorate on Brexit and how it was possible there to thwart the result, and to use technology and algorithms, etc., to affect the outcome in a negative way. There are dictatorships around the globe that have an anti-western and antidemocratic agenda and would want to damage the EU and the western sphere of influence. For that reason, it is particularly serious and I believe Brexit was a very clear example of that.
What is at the heart of this motion is a desire that there be governmental regulation and control; that there be supranational control through the EU and international bodies; that there be control at domestic government level; that we monitor the threats there; that we ensure ethical practice; and that big technology business does not write its own agenda here. I know from private conversations with Senator Malcolm Byrne, when we discussed the setting up of a subcommittee during my time as Leas-Chathaoirleach here - which he knows I was very favourably disposed towards - that that was an effort to deal with this particular motion today and of our support for it.
Again, of course, as Senator Craughwell constantly warns, we must be more than vigilant in the area of cyberattacks which have the potential to happen. We, in fact, did have one in the health service already. We could have it across social protection, it could hit banking and it could hit a whole myriad of areas. It can be done by one criminal gang or it can be done by a country with criminal or malevolent intent towards this country.
I will conclude on that basis because it is our strong desire to hear the Minister of State's thoughts as to where he see us achieving control, basically. Therein lies the public fear. We are all excited about the good things and want the good things to happen but we are concerned about where it could go wrong. We look forward to the Minister of State's thoughts on that in his response.
Gabhaim buíochas leis an gCathaoirleach Gníomhach agus fearaim fáilte roimh an Aire Stáit. We welcome tonight's motion from the Fianna Fáil Parliamentary Party and, in particular, the calls for a national awareness campaign and the need for more work to be done in the area of AI. Sinn Féin also welcomes the recent Government decision to establish an artificial intelligence advisory council. That is something Sinn Féin has been calling for since earlier this year.
With regard to the AI intelligence council, it needs to deliver a dual mandate of ensuring that we harness the positive powers of AI but, at the same time, protect workers and society from the negative aspects of the technology. It is disappointing that the motion does not have a focus on workers. Greater focus should certainly have been given to the impact that advancements in AI will have on workers, specifically, because it was a key focus of the discussions at the Committee on Enterprise, Trade and Employment.
We know that AI can be a profoundly disruptive technology and can also be accompanied by a great deal of technological anxiety. There is a need for a specific focus on the technological surveillance of workers using artificial intelligence and machine learning and it is essential that the advisory council would include workers' representatives from the trade union movement.
The motion is also disappointing in that it does not address the impact AI and emerging technologies are having on the environment and I would like to focus the remainder of my remarks as the climate justice spokesperson particularly around the toll that AI is taking. This is particularly the case with regard to the exponential growth of data centres. These are colossal facilities, often unseen but omnipresent, and they are voraciously consuming electricity and, therefore, fossil fuels, at an alarming rate.
As we witness the strides in AI, where it is turning industries upside down, we have to acknowledge that there are environmental consequences to this. The energy demands of data centres, as I said, are insatiable. The training of the AI models, particularly those deep-learning models, is computationally intensive. It requires significant computational power and those computers then give rise to excessive heat. Their data centres then require extensive cooling systems to manage the heat generated by the high performance computing, and all of that is adding to their overall environmental impact.
It is interesting that there was an academic paper which was published only a month ago which found that Google's AI alone could consume as much electricity as a country the size of Ireland. Some 29.3 TW per hour is required, and that is just Google's AI, which is a significant increase when compared to Google's historical AI-related energy consumption.
We need to pay attention to the expansion of data centres, to the escalating demand for electricity they have, and to what they are actually doing. We agree with regard to data centres, that moving to remote working and to the digital economy is absolutely essential. Equally, there are elements of AI which will be very beneficial to society. We also have to accept, however, that there are elements of AI which are not beneficial and we need to have greater oversight as to which we want, and as to how much renewable energy we are going to dedicate to facilitating this demand. That is because we cannot just have exponential growth of renewable energy in this country. Yes, we have an untapped resource but we cannot just keep building more and more data centres, and more and more computers for AI, and putting that demand on our renewable energy.
One of the most insidious practices from an energy usage and AI perspective is called real-time bidding, RTB. When you land on a webpage, there is a complex quick and energy consuming option which happens to determine what ad an individual will see. It is gathering information on individuals, about their likes and dislikes, their interests, their vulnerabilities, and even their moods, to then sell their attention to the highest bidder. It was the Irish Council for Civil Liberties that helped to uncover the scale of this practice. The machinery behind the system has a rapacious appetite for energy. Some178 trillion RTB transactions each year in the US and in Europe alone are processed through data centres which use an estimated 200 TW hours of energy each year, which is more than the entire national energy consumption of a medium-sized country. As I said, we accept that AI is with us. It has been with us for a very long time. It has come on in leaps and bounds in recent years. We have to have a conversation about what AI is essential and what AI is not essential and about the point at which we say "Stop" to the energy demand of data centres in this country. If we are just to allow exponential growth of AI and data centres, the Government can tell everybody living in rural Ireland where all the renewable energy will come from, whether it is their coastal communities or on land.
Cuirim fáilte roimh an Aire Stáit. I thank Fianna Fáil for bringing forward this important motion. As the Minister of State knows, the Oireachtas Joint Committee on Enterprise, Trade and Employment has already published a report on the impact of artificial intelligence within enterprise and workplaces and has issued a set of recommendations. That follows on from a series of campaigns by a number of trade unions, in particular the FSU, calling for a particular focus on how employers are using AI for surveillance, management and recruitment within workplaces. I know that previous speakers said this as well, but one of the glaring omissions from the motion is any reference to AI in the workplace. There is no doubt that AI has a crucial role in many other aspects of Irish society but probably none more so than within the workplace.
While AI is very much here with us and will be for a long time to come, the key issue is how we ensure we sufficiently regulate and understand what we are dealing with here. A Microsoft survey was undertaken earlier this year across 700 workplaces, from which we are given to understand that one in five workplaces in Ireland is using AI tools. That goes way beyond what many people expect with regard to the use of AI within employments in Ireland. So much of the conversation about AI is very negative, and rightly so, and there is a lot to be fearful about, but there is a huge amount of benefit to AI in the workplace, particularly in terms of greater efficiencies, work practices, removal of the tedium of certain processes and ensuring more reward in work output. If it is badly regulated, however, as it is now, or if it is only partially regulated by laws that were never designed to regulate AI in the workplace, we run a very real risk of entrenching workplace inequality and having even greater unfair and discriminatory treatment of certain workers.
It is also important to say, with regard to AI in the workplace, that while we have the typical power dynamic between employees and employers, AI brings a whole new dynamic to it now that there is a third actor in the room, and it is not just about another technology tool in the workplace. AI will go on to play a very important role in mediating the relationship between employers and employees. In that regard it is so crucial we get the regulation of this right. I know other speakers have referred to the AI Act we see winding its way through the EU institutions, and it is very welcome that the EU institutions are finally getting their act together and publishing a directive. I know there was a set of ethical guidelines originally produced and they have been now moved into a place where there is a realisation that there must be legislation, which is very welcome. We are very concerned, however, about some of the detail within that Act, notwithstanding the improvements by the European Parliament brought about in the draft Act. Senator Malcolm Byrne has already referred to the concept of self-regulation, that, effectively, the creators of AI will have to designate the risk level associated with their products. To me, the biggest challenge is understanding ex antewhether a product will have a detrimental impact on a worker. Of course, the interest of any creator of an artificial intelligence product or piece of software is to see the benefit in it as opposed to seeing the damage it could potentially do. That is why it is so crucial we have an independent agency regulating artificial intelligence across a whole range of spheres, like what we have in terms of regulating competition in this country or with regard to many of the other regulated sectors. AI has to be seen in that realm.
As to what we can do in Ireland, I have a particular concern that there is an element of the Government sitting back and waiting for the AI Act. I see Senator Malcolm Byrne shaking his head, and that is fine. I am willing to hear what he has to say. From our perspective, however, there is a sense that the AI Act will deliver what is needed. A lot can be done in the interim to ensure we have our health and safety legislation and our employment legislation brought up to standard.
Typically, the retention of data by employers relates to personal information but, of course, that data will now be used in very different ways with regard to management, recruitment and surveillance. There has to be an onus on employers to ensure that they retain accurate data and that workers cannot be sacked for inaccurate data, which effectively is governing, managing or overseeing the work taking place within a workplace.
Similarly, as regards the right to information, the Labour Party brought forward a Bill back in 2021 to stamp out bogus self-employment in this country. We know bogus self-employment happens across many sectors, but we see the emergence now of the gig economy and, in particular, bogus self-employment being the business model, so to speak, within that economy, in particular for the delivery sector, including Deliveroo riders and couriers. Many of them effectively work to an app. There is a key issue now that we can, as a member state of the European Union, move ahead of the AI Act and legislate here to ensure there is right to explanation, a right to information and a right to transparency with regard to the tools in the workplace.
As a member of the media committee, I express my support for Senator Byrne's motion and pay tribute to him for the work he has done in this sphere. I also welcome the Minister of State, Deputy Calleary, and acknowledge and praise him for the work he has done in this sphere, especially in the area of national strategy and support for the companies that operate in this area, for the appointment of an AI ambassador and for the advisory council. They are all notable moves so far and in starting that and exploring their conversation. We spoke earlier at the press briefing for this motion, and Senator Byrne specifically touched on the area of deep fakes and bad actors in electoral contests. I will touch on that area in my contribution as well as the potential for significant distortion in the midst of an electoral campaign.
Last month I acted as an election observer in Poland for the elections there. In the work we did prior to polling day, liaising with our long-term mission observers, the point was made forcefully that the impact on the result was solidified in the weeks building up to the elections. The day of the electoral contest itself was run with high efficiency, but it was in the context of the weeks leading up to it that bad actors in this sphere were distorting the messaging the public were receiving. The number of areas where this is happening now is only multiplying, with the manipulation of messaging, so the point Senator Byrne made about the potential in the whole area of AI for that to add to the confusion is significant. In 2015, Poland ranked 18th in the world as an area where citizens could receive trusted media information across all spheres. By this year, the country had fallen to 66th in the world on that measure. It is a depressing statistic for a member of the EU. Ireland, by the way, is ranked second in the world, out of 180 countries, as a trusted media source, which gives food for thought. We do not want to lose that ranking. There is a very close tie between having a free and trusted media and a free democracy that operates properly as a result. The two are intertwined. We need much more public engagement and media literacy in this whole area. We must make citizens aware of the potential misuse of something that can be very positive.
Senator Craughwell spoke about overregulation and said there is not a need to be delving too deeply into this whole area. The bad forces, he said, would be ahead of any legislation or regulation. We have been here before with the Online Safety and Media Regulation Act 2022, which we spent a long time on between Committee Stage and discussions in both Chambers. The point of that legislation was to empower a commissioner and an office to deal with emerging technologies. That was the whole point. It was not just about setting down a list of prescriptions in legislation and saying they are the only areas within which the regulators can operate. The purpose of the legislation was to empower a commissioner to deal with something that is happening at a rapid speed. The same applies in this particular sphere as well.
Senator Sherlock spoke about efficiency and work practices and the benefits of AI for employees as well as employers. That is to be welcomed. Senator Joe O'Reilly spoke about the positive impact of AI in the whole area of health. We need to acknowledge all of those combined benefits. Equally, in the context of this motion, we must acknowledge the challenges for our companies, our electorate and, most particularly, for our young people. I have a deep fear regarding the whole area of media literacy; there is a need for public engagement in that regard. I acknowledge the many good things that arise from these technologies. Indeed, I have discovered that it is possible to write love poetry in two seconds using artificial intelligence. For all the Romeos out there who want to write something quickly in the build-up to Christmas, they can make a lovely love poem in two seconds.
This has been a worthwhile debate. I pay tribute to my fellow media committee member, Senator Malcolm Byrne, for bringing forward the motion.
Gabhaim buíochas leis an Seanadóir Malcolm Byrne as ucht an rún seo faoi chúrsaí intleachta saorga chur os ár gcomhair. Bhí díospóireacht shuimiúil againn agus gabhaim buíochas le gach Seanadóir a bhí páirteach inti. I thank Senators for this useful debate.
The national Al strategy, Al - Here for Good, has been in place for two years. I acknowledge the work that was put into the strategy by my predecessor, Deputy Troy, who was working on it long before there was a lot of attention on the issue. I published a progress report on 9 August this year that highlighted some of the achievements Senators have acknowledged this evening. These achievements include the appointment and excellent ongoing work of our national Al ambassador, Dr. Patricia Scanlon, and the establishment of an enterprise digital advisory forum. The first iteration of that forum has now finished its terms of office and I will be appointing a new forum shortly. We have had the publication by the National Standards Authority of Ireland of its AI Standards & Assurance Roadmap. We have also established a number of hubs, including Ireland's European digital innovation hub for Al at CeADAR in UCD.
This motion recognises the positive environment the Government has created for the development and deployment of trustworthy, person-centred Al. I absolutely recognise the concerns people have, including those expressed in the motion and the debate. It is important that they are articulated and discussed. Regarding skills development, priorities across government for the year ahead include supporting workers, businesses and enterprise on Al adoption, as well as upskilling, reskilling and adapting. The Department of Further and Higher Education, Research, Innovation and Science, the Department of Education and their partners are working hard on the skills dimension of the national digital strategy, which sets out two overall targets to support all cohorts of society to benefit from digital transition. The targets also seek to address skills challenges that are being faced and will be faced by workforces in the face of automation risks and emerging technologies, including Al. Senators Boyhan, Sherlock and Cassells referred to that. Work continues to focus on digital skills provision, the identification of potential digital skills gaps and future skills needs, and progressing the digital strategy for schools and the adult literacy for life strategy. The OECD review of Ireland's skills strategy for 2022 addresses the challenges for our society in regard to sustainable skills development in the face of digital transition and emerging megatrends, including artificial intelligence.
Regulation is and will continue to be a significant part of our approach. Influencing the regulation of Al in Ireland and internationally, including, in particular, the development of the EU AI Act, is one of my priorities. I assure Senator Sherlock that we are not sitting back and waiting for it to happen; we are part of making it happen. I am slightly confused by Senator Craughwell's approach. He stated that countries and governments are one step behind companies and, therefore, we should not regulate. If we do not regulate, however, we will be 101 steps behind companies. Now is the time to regulate and put the guardrails in place. That is relevant to some of the issues referred to by Senator Byrne.
The EU AI Act sets out harmonised rules for the development, placement on the market and use of artificial intelligence systems in the Union. The proposed regulation aims to ensure the protection of fundamental rights and user safety and to develop trust in the development and uptake of Al, thereby enhancing EU investment and, crucially, EU innovation. The trilogue negotiations commenced in June this year. There are a number of complex technical and political points that are currently being negotiated. I acknowledge the huge work my officials are putting into these negotiations. Progress is being made and the regulation is expected to be agreed, at the very latest, in spring 2024. Following that, it is expected that the application of the regulation will be by way of a phased approach that will commence 18 months after its entry into force. The operational and practical impacts of the regulation will depend on the final outcome of the negotiations. My officials are currently monitoring developments with a view to drafting a regulatory impact assessment of the implementation and enforcement framework for the Act, to ensure the forthcoming regulation will be implemented coherently, with functions being assigned to appropriate bodies.
Of course, the EU is not the only international forum in which Al is currently being considered. As Senator Byrne acknowledged, I attended the Al Safety Summit 2023, which took place in Bletchley Park, London, on 1 and 2 November, on the invitation of the UK Secretary of State for Science, Innovation and Technology, Michelle Donelan. Ireland was one of only six EU countries invited to take part. All countries attending the summit agreed to the Bletchley declaration on Al safety, which recognises a shared consensus on the opportunities and risks of particular uses of Al and the need for collaborative action on frontier Al safety.
Ireland's engagement on Al development, in particular in negotiating the EU AI Act but also in other international forums looking at Al governance, emphasises the importance of ethical, person-centred development and deployment of Al and promotes an inclusive, multi-stakeholder approach. The more stakeholders that are engaged, the better, more rounded and more realistic our negotiations. During the summit, I participated in a number of discussions involving representatives from across sectors, which reflected the urgent need for a shared international understanding. I participated in a roundtable discussion on the risks from the integration of frontier Al into society. I also participated in a discussion focused on the theme of what the international community should do regarding the risks and opportunities of Al. In that session, we discussed where international collaboration is needed both to manage risks and also to realise opportunities from frontier Al, including areas for international research collaborations. The summit provided the opportunity for informal engagements across all of the sectors whose active participation will be required for both the economy and society to realise fully the enormous opportunities that present, but also to mitigate the risks presented by deployment of Al. The participants included representatives of governments, non-governmental organisations, industry, academia and civil society.Ireland joined the Global Partnership on Al, as has been referred by Senator Joe O'Reilly, in November 2021. The GPAI is built around a shared commitment to the OECD Recommendation on Artificial Intelligence. The GPAI brings together engaged expertise from science, industry, civil society, governments, international organisations and academia. Ireland has nominated three Al experts to participate in research work being undertaken by GPAI.
My officials, as has been referred to by the leader of the delegation, Senator O'Loughlin, and I acknowledge her work along with that of Senators O'Reilly and Boylan on the Council of Europe, are actively involved in negotiations on the council's legal convention on Al. Ireland supports the balanced approach being taken by the committee of the Council of Europe on Al, which is based on human rights, democracy and rule of law, and it strives to ensure ethical and trustworthy Al where the rights and safety of the individual are protected while remaining conducive to responsible innovation. I understand that the convention is expected to be finalised in March 2024 and adopted in May 2024. It will then be open to both members and non-members of the Council of Europe to ratify the convention and adopt it at a national level. Collectively, all of these guardrails are designed to ensure that there is trust in Al which, in turn, will support ongoing, responsible and ambitious innovation in the AI area.
A number of Senators have expressed concern and are right to do so about electoral integrity. In 2024, more than 2 billion voters across 50 countries, including the United States, the European Union and India will head to the polls. We have seen in other jurisdictions that democracy is under threat, not from the technology but from malicious actors who misuse it. We must do everything we can to protect and preserve our electoral integrity in Ireland. The experience of Senator Cassells and other election monitors is useful in this regard.
As Senators will know, the Electoral Commission has been established and is now operational. The Electoral Commission will also promote public awareness of misinformation, disinformation and manipulative or inauthentic behaviour. It will also establish educational and information programmes to help address this issue. I have met with the head of the Electoral Commission to discuss AI and I am satisfied that he is engaged in this space. The Minister for Housing, Local Government and Heritage, Deputy Darragh O'Brien, is committed to putting in place a robust regime to provide for transparency in political advertising and protect our elections against malign or hidden influence. While not yet commenced, Parts 4 and 5 of the Electoral Reform Act 2022 provide for these matters. Part 4 provides a regulatory framework to ensure there is transparency in respect of online political advertisements, which are purchased for placement, display, promotion or dissemination during electoral periods. Part 5 provides An Coimisiún Toghcháin with monitoring and investigatory functions to prevent the following: the dissemination of disinformation relating to online electoral information; misinformation relating to online electoral process information; and the use of manipulative or inauthentic behaviours online during electoral campaigns.
I am conscious that speed will be of the essence during election campaigns. This is an issue that I have discussed with the head of An Coimisiún Toghcháin and I know that it is an issue the commission has taken on board. Even in advance of AI we all know that deep fakes are in operation during political campaigns and, indeed, online political campaigns. An Coimisiún is engaging with actors both
internationally and in Ireland on effective and best practice tools to counter misinformation. I encourage Senators with experience of this to engage directly with the commission.
AI is a hugely important area. To support the national AI strategy I am establishing an Al advisory council. The council's role will be to provide independent expert advice to Government on artificial intelligence policy, with a specific focus on building public trust and, accordingly, promoting the development of person-centred Al. Its first role will be to provide expert guidance, insights and recommendations in response to specific requests from Government. Its second role will be to develop and deliver its own work plan of advice to Government on issues in artificial intelligence policy, and provide insights on trends, opportunities and challenges. Its third role will be to engage in public communications aimed at demystifying and promoting trustworthy, ethical and person-centred Al and that will involved a public awareness campaign of AI during 2024.
I thank Senator Byrne for tabling this motion. I wish to conclude by noting that Al technologies will impact all parts of society and the economy, not just in enterprise but, equally, in health, housing, agriculture, defence and right across this sphere. AI will change employment patterns but we also need to be ambitious in how we engage with the technology. More detailed awareness of the risks of AI is required and more detailed and understanding of the opportunities across all stakeholders.
It is important to be mindful that while Al is a pressing focus tonight and, indeed, right nowthere are a range of other technologies like quantum computing and biotechnologies that will have similarly transformative impacts on our world. All of us, in both Houses, must be ready to inform, education and challenge ourselves on the technological changes that are under way. We must also strive to communicate not only the risks but the opportunities of these technologies to the public. Our leadership on this will set Ireland up to make sure the values that this Republic holds dear are preserved and protected while adapting to and harnessing the technological changes that are happening before our eyes and will happen for the common good of our citizens.
I thank the Minister of State for his comprehensive response. I also thank my colleagues here in the Seanad for their input on this issue. A number of people have praised the role of the AI ambassador, Dr. Patricia Scanlon. It might be in order at some stage, and perhaps in the new year, for us to invite her to address the Seanad and inform us about her work, and her perspective as to how AI is impacting on the ground.
I agree with the Minister of State that my motion focuses on AI but the full title was artificial intelligence and emerging technologies. When we see the convergence of those new technologies then that is how we are going to see all of our lives change. The Minister of State did make reference to the emerging global mega trends and how we will respond to those. In my motion I also suggested that we look at establishing an Oireachtas committee of the future, and such a committee has been established in other parliaments, most notably in Finland. A committee would allow us to look at some of those bigger picture issues in a cross-party way and bring in external experts to explore those issues.
I welcome the Minister of State's support for Parts 4 and 5 of the Electoral Reform Act. I agree that speed is of the essence because, as everybody knows, we will face a number of elections within the next 18 months so it is critical that we have the necessary safeguards in place.
I disagree with comments made earlier because I believe that the Minister of State has been very proactive of the AI Act. He has said that he optimistic that the legislation would be agreed by the spring but it is critical that the Act is agreed within the mandate of the EU Commission, and obviously before the new Commission and European Parliament elections next year when we have the findings about how that impacts. That is why it is also important that we have a special or select committee, whether it is of this House or the Houses jointly, to look at the impact of that legislation on all aspects of Irish society and, indeed, to have public hearings in that regard.
I strongly agree with Senator Marie Sherlock on the impact that AI will have in the workplace, particularly as we know already that companies use AI especially in the space of recruitment. Therefore, we have got to ensure, in whatever way AI is deployed, that we guard against algorithmic bias and that means right at design phase that we build in the necessary safeguards. AI is not just used for recruitment. It is used for retention practices and a whole range of areas. Therefore, we should have a public information campaign.
I welcome the Minister of State's commitment to continued investment in research because Ireland is and can be a real leader in AI research both in terms of the development of new AI and other technologies but, equally, on the impact of AI on society. The important role of social scientists in looking at both the potential impact and how we should respond is critical.I am glad the Minister of State acknowledged the super work being done by CeADAR out in UCD. I also mention ADAPT, the SFI research centre operating out of Trinity College involving UCD, DCU and quite a number of the other universities. Our higher education sector is very willing and ready to respond.
Senator Boylan rightly raised some of the concerns around technology and its carbon footprint. We must also remember that artificial intelligence can play a big role in addressing this existential crisis around climate change. By using AI effectively, we can look at ways to efficiently allocate resources, optimise supply chains to ensure we can promote sustainable manufacturing processes and ensure we use our energy in the most efficient and environmentally-friendly way possible. This is also why we need to invest in research in much of this space.
I notice Senator Sherlock picked up on this as well, with Senator Craughwell, in respect of being cautious around some of the regulation and legislation. We must set out the guiding principles in this regard. We cannot legislate for every inevitability, but it is critical, particularly for democracies, that we put these guardrails in place. I am grateful to the Minister of State for providing the leadership on this issue and for the contributions of everyone in the Chamber. I hope we will continue to lead the public debate around AI and, indeed, other emerging technologies.