As if it were the launch of a best-selling video game, the Government keeps secret what it is going to do with artificial intelligence. We do know what the main plot lines are and who will be behind the development, but little else. There is no specific start-up date and it is not known what will be the (financial) requirements that will be necessary to move this entire project. The scientific community is running out of nails from so much hype.

What is known right now is that in the fall of 2017 the government called a group of experts to develop a strategy on AI and big data. From what it was about then, a strategy was created that evolved for months until it was settled into something serious, ready for publication. But the motion of censure against Rajoy in June 2018 caused the portfolios to change and it was necessary to take a few steps back in the development of the strategy. When it seemed that the project was going to be cornered (not having the general state budgets approved helped), the European Commission raised its voice and forced the EU countries to define their plans: before the summer of 2019, all countries should have their national strategies on R & D & I in AI presented. Spain did its homework and delivered the document in March.

Did this mean that the lines of how Spain was going to develop AI were set? No. As is often the case when homework is turned in early, the content rarely goes into detail. The ‘Spanish Strategy in R + D + I in Artificial Intelligence’, 48 pages with a foreword by Pedro Duque, outlines the framework – fundamentally theoretical and with an abundance of verbs in the future – of how the development of AI should be in our country , but leaves open how it will be done, when and how much it will cost, chapters that should appear in a new document called ‘National Strategy for AI’ that should come out soon. In the absence of the national document, some autonomous communities already have theirs: Catalonia and Valencia, for example.

The big problem of AI in Spain

Although it is undeniable that AI is beginning to gain importance for the government (as shown by the appointment of Carme Artigas as Secretary of State for Digitization and Artificial Intelligence), the truth is that Spain has been carrying well-known problems for years. Spanish researchers denounce that there is no transfer between the knowledge that occurs in universities and companies. “We are fine if we look at the scientific production, the articles published and indexed in SCOPUS and the impact of those articles ”, explains Nuria Oliver, doctor in AI from MIT and participant in that group of experts in 2017.“ It is true that we have two powerful sources of innovation, which are Madrid and Barcelona, ​​but we failed to adopt the use of AI in the existing business fabric ”.

Professor and researcher Ramón López de Mántaras, founder and former director of the Artificial Intelligence Research Institute of the CSIC (IIA-CSIC), who was also part of the 2017 group of experts, also recognizes this duality. The professor places Spain within the European top 10 when it comes to AI research. Countries like the United Kingdom – “the number one by far” -, Germany, France, Switzerland, Israel, the Netherlands, Finland or Italy are better positioned. If we open the classification to the whole world and the United States, Canada, China or Japan enter, “then we may have about 15 countries ahead, but not many more. In terms of research, Spain occupies a traditionally dignified position. What does not work is the transmission of all that research to the industrial fabric. We have many start-ups that are powerful and do it very well, but they are not international benchmarks ”.

How can it be explained that if knowledge exists and it is often transformed into spin-offs and start-ups more than competent, a solid industrial fabric specialized in AI does not develop in Spain? Francisco J. Martín, co-founder of BigML, a company specialized in machine learning based in Oregon, United States, and an office in Valencia, points to both the public and private sectors. To the public sector: “There is no agile program of financing with public money to small companies as there is in the United States.”. And to the private sector: “the large Spanish companies, which are the ones that should lead investment in AI, do not invest in technology locally by hiring start-ups of two or three youngsters to develop their project, but rather all their investment technology is going to be bought from the big North Americans like IBM, Google or Microsoft. They do not help in the least to build a local technological network, and instead they try to be innovative by sending their managers to Silicon Valley, as if the innovation were to come to them by osmosis”, He concludes.

The main demands

The main demand of the scientific community does not admit surprises: more money is needed (remember that it is not yet known how much money will be allocated to AI, when countries such as France, the United Kingdom, Germany or Finland have already published their figures). “But not everything is money,” explains López de Mántaras. From the point of view of universities and research centers, incentives are needed so that experts 1) do not leave and 2) if they did, they can return. López de Mántaras, recalls that in his last eight years at the helm of the IIA-CSIC he has seen more than 30 people march (and that is about 60 researchers). “Now they are at Amazon, Google, Apple, Oxford University… Not one of them has been able to return and it is not because they do not want to, it is because they do not have the opportunity to do so. It is very sad and I think that in this sense we cannot say that we are better today than a decade ago. We have let go of a lot of talent that has not returned. “

For Nuria Oliver, the Government should look towards the business fabric and try to “inspire and develop an ecosystem of technology-based companies that are not restricted to physical territory, but can operate outside our borders, but from Spain.” existing companies, “we should help to identify which elements can be digitized to make companies more competitive thanks to AI. It is true that we have many SMEs, but in other European countries as well, and public-private collaboration and investment agreements are being made to encourage digital transformation ”, he emphasizes.

The public sector also has the challenge of modernizing itself, says the expert. “The Administration is still a great company, and there is a great opportunity to adopt AI both to optimize processes and to offer more personalized services to citizens. It has a double role: to develop AI within the institution itself and also to apply it to the provision of services that the administration offers to citizens ”, Oliver concludes.

An AI made in Europe

It’s no secret that all this AI rush comes directly from above and from the fear that the European Union will move away from the race for AI led by the United States and China. It is estimated that the EU will spend between 2021 and 2027 more than 100,000 million euros from public and private sources through the Horizon Europe program, an amount that will be added to the 20,000 million that were injected between 2018 and 2020, to the 700 million destined to research or to the more than 7,000 million that come from the European Digital Program, among other equally notable amounts. And we are talking about investment only for AI. However, the difference between what Asia or North America invest with respect to Europe is between 3,000 and 15,000 million euros per year, says Oliver.

The EU plans to spend more than € 100 billion on AI alone between 2021 and 2027

At the academic and research level there are also initiatives to add muscle at the community level. López de Mántaras is part of the Confederation of Laboratories for Artificial Intelligence Research in Europe (CLAIRE), a pressure group that is committed to “developing AI for living and that rejects any military use or that is not socially responsible and impactful positive ”, explains López de Mántaras. Nuria Oliver, for her part, is a member of the board of trustees of the European Laboratory for Learning and Intelligent Systems (ELLIS). “With ELLIS we not only seek to promote excellent research in machine learning, but also to have a positive economic and social impact. We also want the next generation of talent to stay in Europe, because we know that great talent is not in Europe and if it is, it does not work for European institutions”Emphasizes Oliver, who recalls that the brand-new 2018 Turing award winners (Yann LeCun, Yoshua Bengio and Geoffrey Hinton), two French by origin and one British, do not work in Europe, but in North America.

“I don’t think the race against the United States and China is served”

For Oliver, with all these initiatives, Europe is planning what the medium-term future of AI will be like across the continent, and it is doing it in its own way. “In Europe we have a much greater focus than in other regions on ethical and regulatory aspects. We were pioneers with the GDPR and now they are following us in other parts of the world. I think it is a good opportunity for we Europeans define how we want to develop AI beyond the technological element, and we take into account four dimensions: ethics, legal and regulatory, social and economic and labor. ”

“I do not believe that the race against the United States and China is served, and I want to believe that the European vision will allow us to develop AI made in Europe that facilitates both innovation and the preservation of the quality of life in Europe”, Oliver concludes. The approaches that Spain makes regarding the development of this technology also go in the same direction (there is no other possibility, the EU is in charge). How exactly it will achieve this, and if it will meet the demands of scientists and entrepreneurs, we will find out in the coming months.

Images | Home |