Speech by the President of the Government of Spain at the closing ceremony of the event 'IA360: a plan for the responsible deployment of AI'

SPEECH BY THE PRESIDENT OF THE GOVERNMENT OF SPAIN, PEDRO SÁNCHEZ

Good morning, first of all to my colleagues in the Council of Ministers.

Secondly, to the representatives of business and, of course, the trade unions as well.

Thank you for making the effort to come here on a Monday morning, to your home, the Moncloa Palace, the seat of the Presidency of the Government of Spain, to discuss a subject as important and as far-reaching for our societies as artificial intelligence - and, as the Vice-president rightly pointed out earlier, and, of course, as our distinguished panellists have also highlighted, it is a challenge. It is a debate that is increasingly coming to the fore in conversation - not only in public discourse, but also in public discourse - a conversation in which Spain has, incidentally, participated directly or indirectly all along.

As the United Nations General Assembly gets underway this week, it is important to remember that last year, the United Nations General Assembly approved the creation, for the first time, of the Independent International Scientific Panel on Artificial Intelligence, modelled on the Intergovernmental Panel on Climate Change, precisely because we are very aware of the need to establish minimum standards across all nations for the development of this significant technology. Spain has played an active role in its inception and even hosted the first meeting, in the Constitutional Chamber of the General Courts, because we specifically wanted to convey the message that it is not only science, and experts, but also political power as expressed in a democracy - logically, where national sovereignty resides, namely the General Courts, and where legislation is passed - for this is the place where we must achieve that intersection, that convergence between science, technology and, of course, the aforementioned regulation.

It is clear - and I don't think we need to hear from those working within these companies - that this challenge posed by artificial intelligence is no longer something for five, ten or 50 years' time. It's happening right now, at this very moment, and this also gives the impression that we're a bit late to the party, when it has previously been said that we need to bring it back to a human scale, above all so that we can understand it, take it on board and actually make the very best possible use of it.

I'd like to start with an anecdote that some of you may already be familiar with: at the end of the 19th century, the United States was experiencing what came to be known as the dawn of the Gilded Age, a period in which vast fortunes were amassed by a handful of tycoons. And one of them was William Vanderbilt, who - for those of you who don't know him - was the great railway magnate. And he is credited with the following words, which I shall quote below: "To hell with people. I only work for my shareholders". This is what he said at the end of the 19th century. I believe that phrase symbolises many things, but what I wanted to say is that it exemplified, like few others, the ethos of that era: the arrogance of a class that believed itself to be above everything and everyone. The only rules he would accept were those he imposed on himself; never those laid down by the public authorities, for whom - incidentally - he made it clear with this statement that he had no regard.

But if there is one thing that history shows - and we are seeing this once again in the development of artificial intelligence, though not only there - it is that self-regulation does not work. That the incentives of the few who control cutting-edge industries - I recently read an interview with one of these leading tech-oligarchs in which he said, 'Trust us'; well, I'm afraid not; let's all trust one another to decide what we can do with this technology-what I was saying is that the incentives of the few who control these cutting-edge industries will never be aligned with the public interest, however much they try to convince us otherwise, and that, therefore, a transformation on the scale of artificial intelligence requires clear rules, decided by everyone, controlled by everyone and aligned with the values and principles that humanity stands for.

Therefore, if artificial intelligence remains concentrated in just a few companies and a handful of countries, its impact on shared prosperity - and, consequently, on the very idea of democracy - will be extremely negative, if not devastating. That is why I believe we must act urgently, because unregulated, opaque artificial intelligence - running amok, as I have heard it described; I really liked that interpretation of it as not a race, but a stampede - and concentrated in the hands of a very few is the perfect recipe for disaster.

Artificial intelligence without the guidance of collective human intelligence is not intelligence: it is chaos and stupidity. And if we act now - and do it right - I think there's a silver lining. And here, too, there has been discussion of the benefits and opportunities offered by artificial intelligence. Developing responsible artificial intelligence, as has been discussed here, and as is also expressed in the title of this plan we are presenting today - artificial intelligence with a humanist perspective and vision geared towards collective well-being and prosperity, enabling us to build fairer societies - I believe this must be in everyone's interest - not only that of democratic governments, but also, naturally, that of civil society as a whole, represented here today by business leaders, workers and, ultimately, a significant part of our country's organised civil society.

We therefore need supervised artificial intelligence that is traceable, which helps us make the best decisions, so that it does not replace human judgement where the latter must prevail. And there are countless other cases which I need not go into at the moment, but which are clearly on everyone's mind.

This technology can bridge many divides: social divides, economic divides, geographical divides… Or it can do the opposite: widen them even further. It can create new jobs. Or it can destroy more than it is capable of creating. It can sustain its growth through the rational use of natural resources - we have spoken earlier about energy, but when discussing data centres, we should also bear in mind resources as vital and scarce as water - or contribute to their depletion at a crucial juncture for mitigating and adapting to the other major transformation we must undertake to tackle the challenge of the climate emergency, namely the energy transition and the ecological transition. It can provide our teachers with new tools and broaden our students' horizons. Or simply offer everyone shortcuts that undermine that learning process - which is clearly one of the warnings, one of the red flags, that experts are always reminding us about.

So, what I mean is that this future is not set in stone, nor is it determined by the forces that are now frightening even their own creators. That future depends on all of us. We'll write it ourselves. And if we do it right, artificial intelligence can democratise access to these new technologies, driving productivity improvements so that they reach all sectors, all workers and every household in our country more quickly. If we get it right, I repeat, AI can make workers more valuable to their companies, not more dispensable. And if we get it right, artificial intelligence will create opportunities in more sectors than it will put at risk.

But, I must stress, we have to get it right. We have to do it properly. Because artificial intelligence has, in the space of four years, covered the ground that other general-purpose technologies took decades to cover. And this is also what, in a way, is worrying and frightening. The truly exponential rate at which this technology is advancing. And that, therefore, raises a crucial debate that democratic societies such as ours must address. And the question is not whether this technology should be developed - that, of course, is not up for debate. We can't hold back the tide. What is under discussion, however, is the conditions under which this artificial intelligence and this technology should be developed.

And that is where I link in with what we are presenting today: the AI 360 Plan, for the responsible deployment of artificial intelligence, which is what we are presenting today. A roadmap setting out specific actions, with specific milestones and specific dates, to be implemented over the next 12 months. A plan to enable our country to capitalise on the opportunities that artificial intelligence offers our economy - but not just our economy. And it also places the focus on public safety, trust and the protection of the most vulnerable, starting with children.

I believe - and this has already been mentioned here by the Secretary of State and also by the Vice-president, and I am often told this by all the ministers, including, of course, the minister responsible for this area, Óscar López - that we are starting from a solid position. Almost half of Spain's working-age population already uses generative artificial intelligence, which therefore places us amongst the leaders in this field. This was shown in the video - 7th place. Here in the notes, it says our country's ranked 6th in the Stanford University ranking. It doesn't matter whether it's 6th or 7th. The important thing is that we are in the top ten when it comes to the use of generative artificial intelligence. We rank second - and this is also very important - in terms of graduates from Higher Vocational Training in Communication Technologies within the ICT sector. We have highly competitive, robust digital infrastructure and computing capacity, as well as competitive renewable energy, which we must continue to develop.

But, furthermore, I believe that what is important is that our country - not just the public authorities, but also a large part of our country's civil society - has understood from the outset that innovation and rights are not divergent paths. That we can develop cutting-edge technology, such as artificial intelligence, whilst at the same time demanding that it be safe, transparent and respectful of people. And that is why we are ranked second in the world for artificial intelligence governance, according to one of the leading rankings, namely that of Stanford University. As the Secretary of State mentioned earlier, we were the first country to set up the Artificial Intelligence Supervisory Agency. Incidentally, as part of that decentralisation policy, the Agency is now based in the Galician city of A Coruña. We are promoting the United Nations Laboratory in Valencia. And - this is also significant - under the Spanish presidency of the European Union, the first major agreement was reached on a European Artificial Intelligence Act, the world's first comprehensive legislation of its kind, which sparked debate and controversy but marked the first step towards supranational regulation of this technology.

Few countries, therefore, I believe, have been as clear in their defence of a humanist-inspired artificial intelligence as Spain is today. And I also believe that time is proving us right, because the plan we are presenting today is built on those principles and that experience, and is also driven by an ambition to continue expanding it. For example, in a few months' time at the Ibero-American Summit, also at the Ibero-American level, as Vice-president Carlos Cuerpo mentioned earlier.

This plan is based on a number of key objectives, which I shall summarise in four points, in an effort to be as brief as possible. Firstly, we need to strengthen our technological capabilities. This is clear. I have previously spoken about data, and computing power has also been discussed. We have strategic projects such as the artificial intelligence gigafactory, and we are working with the Barcelona Supercomputing Center to develop new artificial intelligence models for use in key areas, such as the fight against the climate crisis, healthcare and energy. Committed to the development of data centres that meet the highest energy and environmental standards. Because we must also develop this intelligence, not by turning our backs on the local communities who live in these territories and near these data centres, but by ensuring that they, too, see it as an opportunity in their local areas and not as a means of depleting their scarce resources. Therefore, our autonomy over the data stored there must be guaranteed and, of course, our natural resources must be respected. Firstly, therefore, as I said earlier, that technological muscle.

The second is to turn this rapid social adoption of the technology into an economic driver and also, as mentioned earlier, into a breeding ground for talent.

There are some very interesting initiatives, such as the Artificial Intelligence Voucher for SMEs and the self-employed, which has a highly ambitious objective. The fact is that by 2030, more than half of our companies will have integrated artificial intelligence into their production processes - and, ultimately, into their business operations. To achieve this. This has already been mentioned here by some of the panellists. We must make a firm commitment to talent and its development.

We are going to adapt the secondary school and vocational training curricula so that our young people learn to use AI and to interact with it in an intelligent and responsible manner, but also to ensure that the roll-out of AI does not undermine the learning process - which is, of course, one of its main risks. Because, clearly, AI presents opportunities, but also challenges and threats that we will have to regulate and control.

Thirdly, to establish a framework for the governance of artificial intelligence that enables us to tackle these challenges. I believe that the risks posed by uncontrolled AI are becoming increasingly apparent. We're seeing this in many news reports about cybersecurity issues arising from various applications and models. This has also been confirmed by recent statements from some of the key players, although in the end it is hard to know how much truth there is in them. But there is also likely to be a strategy behind the search for and recruitment of investors by some of these large companies. In short, whatever the case may be, what is clear is that security must also be an absolute priority in this plan. And we're going to tackle it on two very specific levels. Firstly, by strengthening our cybersecurity defences. And, secondly, by stepping up monitoring of the most aggressive frontier models.

The fourth and final objective of the plan we are presenting today is probably the one that gives meaning to all the measures it comprises, and it is the one I would like to see go beyond the formal presentation of this plan. The fact is that we must all work together - the private sector, workers and the public sector - to promote a new Social Contract for Artificial Intelligence. A new Social Contract for Artificial Intelligence - in short, a major national agreement between employers, workers (and therefore trade unions), employers' organisations, parties and public authorities; a major agreement designed to provide people with security in the face of the changes to work and life that this technology will bring about. And that is why next month - I can now announce - we will be bringing together the social partners to make headway on this major national agreement, this new social contract for the country.

Friends, ladies and gentlemen, as has been said earlier, I believe we have talent; we have talent in abundance among our young people, our scientists - both men and women - and our mathematicians. Just a few days ago, I met with the Minister for Science, Research and Universities and some of these mathematicians at the Autonomous University. So, we have the talent, we have the companies, and I believe we also have a dynamic start-up ecosystem. We have connectivity, we have renewable energy, and we are building the infrastructure needed to advance artificial intelligence in a responsible manner. But, above all, I believe that we have a certain moral authority as a country to drive forward a bold agenda in this regard, not only here but also at a global level, so that we are not at the mercy of the monopolistic - and indeed abusive - practices of these large oligarchies.

An agenda that holds companies legally and financially accountable - as is the case in many other sectors, some of which are represented here - for the harm caused by their activities; and therefore for the harm caused by their algorithms, whether in the form of deepfakes, mental health issues affecting our children and young people, or the theft of intellectual and industrial property.

We have many multilateral mechanisms to ensure that cutting-edge AI serves humanity, rather than the other way round. We have, for example, the European Artificial Intelligence Office; we have, as I said earlier, the United Nations Scientific Panel; we also have global dialogue between governments and businesses on artificial intelligence; and in all of these, our country is a leading player; it is not a mere bystander, but an active participant, one that engages, puts forward proposals and, ultimately, seeks the responsible development of this important technology on a global scale; we are therefore striving to build consensus and foster commitments amongst different societies.

We shall therefore redouble our efforts to ensure that these mechanisms move forward, working alongside other like-minded countries. I believe there are numerous examples, whether in Canada, Japan or within the European Union, of countries that are clearly aligned with our approach to the responsible development of artificial intelligence. Ultimately, we are doing this because no single state can, on its own, regulate a technology which, as has been said, knows no borders.

I would like to conclude once again by thanking the many leading figures from a wide range of sectors who view artificial intelligence as an opportunity, but also as a tool that we must regulate. And, in addition to expressing my gratitude, I would like to conclude with another historical reference. Today, just as 150 years ago, some business magnates continue to argue that regulation is the enemy of innovation - some thoughts on this matter have been expressed previously. They are therefore wary of the state because they believe they are above it and think that there should be regulation - regulation which they themselves must dictate. And believe me, when I have had the opportunity to speak with these tech-oligarchs - some of them, at least - I found that, now that artificial intelligence has been unleashed, what they want is to create barriers to entry to consolidate precisely that monopoly or oligopoly.

In short, they couldn't be more wrong. It would be good if today's oligarchs did not make the same mistake as the tycoons of the past. Because, at the end of the day, no one is above the public interest. And believe me, in this field I think there are a thousand reasons why that interest should prevail. Perhaps the most important point of all is that nowhere is it written that technological progress guarantees social progress, as Vice-president Carlos Cuerpo asked rhetorically earlier in his speech. It is up to us to make this happen, and that - and nothing else - is the purpose of this strategy, of the Plan we are presenting today: To promote a new agenda for the development of artificial intelligence, underpinned by a new social contract that places the people of our country at the forefront - people who deserve to be at the heart of every public policy priority, as well as the collective national effort we so urgently need in the development of artificial intelligence. We're counting on all of you.

That is all, thank you very much.

(Transcript edited by the State Secretariat for Communication)