Photo: The Secretary of State for Digitization and Artificial Intelligence, Carme Artigas, received us at Red.es’ headquarters. Credits: Courtesy of the interviewee
By Marta del Amo
Rather than building a house from the ground up, the Secretary of State for Digitization and Artificial Intelligence, Carmes Artigas, is arguably building it all at the same time. Aware that Spain is lagging behind in some of the key digitization areas, but also that the country must begin to move forward in tackling the major challenges of the future, her organization is launching all kinds of projects to address challenges as diverse as digital literacy and pilot testing Europe’s AI regulation.
Your Secretary of State has just launched the R&D Missions in Artificial Intelligence programme to fund strategic projects. Can you comment on the key elements of the initiative?
Spain has a historic opportunity to transform itself digitally, so it is essential to mold our society and production model. In the same way that Moncloa is looking to the future with its Spain 2050 strategy, we are seizing the opportunity by saying: “Which areas of AI should we invest in and start work on today in order to solve the biggest challenges facing future generations?”
The R&D Missions in Artificial Intelligence programme brings together the view of the economist Mariana Mazzucato of the “entrepreneurial state”, but with moonshot thinking typical of Silicon Valley. We have identified five pressing issues that artificial intelligence can help us solve: agriculture, energy design for the 21st century, environment, health and employment. Our goal is to address the Spanish economic model, which is not resilient, so that when GDP falls by 1%, unemployment rises at the same rate, not by 10%.
There are now five tasks, and we will launch five more in the first half of 2022. The first five will be given 10 million euros per task, and subsidies ranging from 65% to 80%. This has never been done before. The only country in the world that has launched an artificial intelligence mission in these terms is Japan. It is a very innovative subject.
This is about grants for large-scale projects involving public-private partnerships. Why do you choose this approach?
Because we do not want it to remain a theoretical R&D exercise. It is a new mechanism to achieve different effects. We want to generate new innovation dynamics in our country, and for that we have to build infrastructure in talent and innovation. We want to help transfer moonshot thinking, a disruptive way of working, to the market.
“I don’t want the way we share data in Spain to be different from what we are planning in Europe, just as I don’t there to be big data silos between departments.”
Sometimes companies won’t collaborate because they don’t have a common goal. The only way to bring about those changes – this transfer of knowledge - and create a talent pool requires everyone to be aligned in the same direction. What gives us hope is that this has already happened spontaneously during the pandemic. The idea is: Why don’t we turn that into a methodology? Why don’t we systematize it?
What we are saying is: I have this long-term challenge as a country, I want to solve it with artificial intelligence and with these restrictions, by 2050. So, we are throwing major challenges at groups of companies and public-private consortiums in order to create ecosystems that can solve them. We require a large company, five SMEs, a research center and a transfer center.
Beyond this economic boost, how can the competition and innovation of Spanish companies be stimulated in a sector where the main asset and constraint is access to data? How will it be ensured that there is enough data to make relevant algorithms and programmes?
To make Spain a leading country in artificial intelligence we are creating data spaces with the GAIA-X project, a European initiative that aims to create industrial data spaces across Europe. Spain has taken the decision to take an active part in its governing board and in the launch of the national GAIA-X hub.
“Spain should take the lead in tourism data in Europe”
We are working with the industry to make Spain a pioneer in the creation of data spaces in some of the sectors in which we have proven to be very strong. The first is tourism. We want to take the lead in tourism data across Europe, just as Germany is in the lead in the automotive industry. I don’t know if each mission is going to correlate with its own data space, but it is a two part parallel project to create standards for data sharing.
Would the idea be to offer data from Europe to these consortiums?
I don’t know. We haven’t really thought about it that way. Our idea is that they discover their difficulties, and we help them to solve them.
Shared data spaces do not imply that we have to wait for Europe to make our national data spaces. What we want is to create an environment to define that model of governance of data spaces at the national level, designed so that they can be shared across Europe. I don’t want the way we share data in Spain to be different from what we are planning in Europe, just as I don’t want there to be big data silos between departments.
Another major problem associated with AI is the underrepresentation of women and minorities in its development, which leads to biases. This situation, coupled with the dearth of female entrepreneurs in Spain, could lead to a sector as biased as those of the leading powers. What options do you see to address this situation?
The issue of biases in artificial intelligence has several aspects: the inherent bias of the training data, which is the technical one; and the bias of the software developer, due to the fact that the work teams are not diverse enough. We must tackle both of these.
On the technical side, the industry is moving and developing technologies and methods to generate synthetic data. We at the National Strategy of Artificial Intelligence, with the creation of a quality seal for algorithms, want to create a methodology for auditing algorithms in Spain, which is already being incorporated in the National Strategy of Artificial Intelligence.
“Digital training in Spain’s big issue is that we have to address citizen’s lack of basic skills, with 43% being digital illiterates, and yet there is also a need to create very advanced profiles at the same time.”
I will tell you one thing in advance: we are finalizing an agreement with the Joint Research Center for Spain to be the pilot country to test Europe’s AI regulation with the industry, to see what works and what does not. Sometimes you make a regulation that is very well written, but then it has nothing to do with what the industry can do. Finding the balance of bringing regulation closer to the industry that has to implement it is precisely that which Spain has offered to lead. As no one else has offered, I am convinced that we are going to do it.
In order to bridge the gap between artificial intelligence and women, we are working on a training plan and also on the ‘Startups Law’, where there is an important section aimed at promoting the creation of digital artificial intelligence companies on women’s behalf.
The big problem of digital training in Spain is that we have to address the lack of basic skills among citizens, where 43% are digital illiterates, and the need to create very advanced profiles at the same time. We are working with universities to improve Postdocs and PhDs, and we are going to launch a scholarship program in the Recovery Plan. There are 300 postdoctoral fellowships in areas related to artificial intelligence, robotics, etc., and we are going to guarantee that 50% will go to women in order to address the deficit.
We also need to address girls’ vocations from primary and secondary school. But I believe that this starting point is wrong: we make children choose between science and literature very early on, and we are not making it clear that all future professions will require soft skills as well as logical and critical thinking. We have resolved this in the National Digital Skills Plan, with Artificial Intelligence scholarships for PhD and Post-Doctoral students, which we will launch next year with a support plan for female entrepreneurs that we will incorporate into something that we have yet to develop, the Spain Talent Hub, a project with which we want to make Spain a hub for international talent.
We are solving some of this through the ‘Startups Law’, and through the nomad visa, but we want to focus on attracting female talent to Spain. We believe that Spain has some high quality-of-life characteristics, including women’s rights – compared with those of, for example, Latin America -, which makes us very well positioned. It is a safe space for women, which can be a major factor in attracting international female talent.
The EU’s draft regulation on AI raised the possibility of companies themselves submitting their algorithms to internal audits, something that was widely criticized. What form do you think the auditing process should take and what kind of variables should be audited? Are you considering the creation of an independent auditing body at the international level?
It is clear to us that we need an independent organization to audit the algorithms. We are going to try to set this up and work on the opportunities and difficulties of this mechanism. I believe that, in the long run, there will have to be quality monitoring systems for algorithms, even if they are national, but connected across Europe, because we share regulations, principles and values.
Spain is a pioneer in putting ethics before technology and in the Charter of Digital Rights itself, where there are specific sections on artificial intelligence and neurotechnologies, with which we are pioneers worldwide. We are also pioneers in avoiding algorithmic discrimination thanks to the Rider Law, which covers the concept of algorithmic transparency.
Even so, many algorithms are bought from foreign companies that we do not know how they work, and it is the user who ends up affected by that algorithm. Just as exports are controlled, should there be some body that controls how foreign algorithms purchased by European companies work?
Here we enter into industrial property issues and intellectual debate, it is not simple. Just as any country makes a credit risk assessment in order to see if it will give you a loan or not, we must understand what the algorithm’s assessment parameters are. The Rider Law states: “if I am an employee then you, my employer, have to assess my performance, but I have to know the criteria used to measure my performance.” And then we’ll get into whether auditing everything is possible or not.
I still don’t dare to say how the algorithms have to be audited. There are experts who are working on it. There are things that can be audited and others that can’t, because there are mechanisms that we don’t know how they work. It’s like saying: “I want to know exactly how the brain works.” We don’t know.
The important thing is that it will have a risk-based focus. If this is good for you and me, maybe we don’t even need to regulate it. If the supplier and customer are happy, and there is no risk of discrimination, why go into that black box.
What we must understand is that you cannot over-regulate something that has yet to be developed, nor can you let technology run free as if it were the Wild West. We are beginning to take the first steps in this difficult balance.
Regarding the Charter of Digital Rights led by your Secretary of State, some experts have lamented its lack of normative force. What will be the Secretary of State’s next steps to promote it?
The charter is like the SDGs [Sustainable Development Goals]. They are not translated into a law, but rather into guiding principles to govern all laws. The Charter of Digital Rights is the realization of those rights that we want to attribute to ourselves as a society and that should inspire all future normative development. We want it to be like the UN principles, something that is permanent and can move forward as technology advances.
That is what this charter aims to do: to have 25 rights that until now remained as a great principle. It is inspiring the European Declaration of Digital Rights, which we have transferred to the European Presidency. We are even collaborating with Chile, on its future Senate, and the U.S. Government has asked us to do so. I am convinced that it will align at an international level, even into a new declaration of Human Rights.
These are the rights that we want to claim as a society, both individual and collective. We do not want to lose the rights we had in an analogue world, but we need to revisit them. How should we reinterpret the right to privacy in a digital world?
I believe that these are steps in the right direction so as not to be spectators of the world to come. The future is not written, we write it every day with our decisions and actions. This is one of them: Spain wants to be a pioneer in these rights and that is what is going to inspire the development of safe laws in this government.
You have stated that you would have liked the ‘Startups Law’, promoted by your Ministry, to have existed when you launched your company? What impact do you expect it to have on Spanish society and its industrial fabric in the short and medium term?
The first thing this law will do is change the playing field. The fact that there is a law for startups offers them specific and differential recognition. It only applies to this group because it has unique characteristics: to grow very fast, to have very important capital injections to manage this growth and to compete globally to capture investment and talent.
Spain did not have mechanisms for this. Issues such as helping the most difficult early stages, increasing deductions by 40%, facilitating the whole issue of visas, international talent, even digital nomads as a result of teleworking, facilitating individual international investments. In addition, the issue of stock options, which we have increased the limit to 45%.
And by no means is the serial entrepreneur being avoided. People have interpreted that you could only be a start-up once. No. That is the big confusion. No, because we are just simplifying the closure of companies that fail, because we believe that they could be a serial entrepreneur.
What is the nuance? When you involve 11 ministries, as we have done with this law, each one sees it from its point of view, and some had much more to say. The Treasury, which is in charge of subsidies, says: “How do I avoid fraud with public money?” They want to avoid that if I set up a startup and I am entitled to subsidies, grants and advantages in the first five years of life (because we have already increased from three to five), what I cannot do is close the company and reopen it under another name, without changing activity, I never make profit and I am like that for 30 years. This is what we want to avoid, it is a question of technical drafting.
I am calm because it makes us very competitive among the countries around us, such as Portugal. We do not want to be a tax haven; Spain has other assets that are much more attractive than the best tax system. Our country’s public services have endured the pandemic and temporary layoffs, and we have a world-class public health system which we don’t want to give up. Let’s not fool ourselves, here taxes have to be paid in order to be able to afford the social welfare we have. Spain will be a paradise for startups, but not only because of its tax system - it has other great assets and attractions.
As much as entrepreneurship and the creation of new businesses is facilitated, the truth is that our economy is lagging behind in terms of digitalization, especially in the field of SMEs. What is your assessment of this situation and how could the government improve it?
This is one of the things that must be reversed as a matter of urgency because they are weighing us down as a country. In the DESI index (Development of the Information Society in Europe) we are in position number 11. We stand out in connectivity and development of digital administration, but we are far behind in digital skills and in digitalization of our SMEs.
An example: during the pandemic, e-commerce has increased exponentially, 50% on average in all sectors and 70% in retail alone, but SMEs in our country have only captured 9% of this market, and this cannot be. The aim of the Digital Agenda is that by the end of 2050 at least 25% of the e-commerce generated in Spain should come from SMEs.
To do that we have to help digitize our SMEs with the SME Digitization Plan, endowed with 4.75 billion euros, aimed at the first adoption process of those small companies that are lagging behind. And before the end of the year, we will launch the digital toolkit, consisting of 3,000 million euros to help companies in our country to use technology to transform themselves.
80% of these SMEs have less than five workers, so we are talking about small enterprises. We need to make sure that they are not left out, that they have an online presence, that they understand what electronic invoicing is and booking appointments online, that they have software as a service and that they have access to customer knowledge. That is the great obsession we have, that Spanish SMEs are not left behind, that they do not hobble us as a country and that we can turn them into digital SMEs.
At this point in the pandemic, given the very low download rate of the Covid Radar app, do you think that it is time to call it a failure? What lessons have you learned from its launch?
It is not a failure of adoption, nor a technological one. Radar Covid exists, it works, and in the autonomous communities that use it, the penetration rate is 18%. The problem is that not all the autonomous communities have decided to use it. Only 50%.
What’s going on here? Big lesson learned: the adoption problem is not technological, but rather cultural. We are a country that has 17 autonomous communities and two autonomous cities, and we have made it available to all of them since September 1, but the decision to use it depends on each healthcare system. Some have done it very well and automatically, while others have not adopted it because they have not considered it appropriate or because they have not been able to provide an emergency response and bring about a cultural change to try to integrate a technological system, in addition to the manual one.
Radar Covid is working, what is not working is the notification that the user receives. There have been 55,000 Covid Radar alerts which means that 150,000 people have been warned that they have been in close contact with someone who has reported a COVID-19 infection. And these 55,000 alerts would be equivalent to 50,000 human contact tracers. We have saved 50,000 human track and traces.
What we have to analyze is why some autonomous communities (Catalonia, Valencia…) with very high COVID-19 incidence have not used it, why they have not been able to organize themselves in order to integrate this system with the manual system.
It is the best possible tool to tackle the higher incidence among young people at the moment. Therefore, we should not write off the application, which has a long way to go. It is an application for the whole of Spain, but difficult to implement in a distributed regional model.
Just one person that you were able to warn three weeks before or 10 days before they were with a high-risk contact and saved their life makes Covid Radar worthwhile. The lesson learned is that we can’t do this if we don’t all use it. Otherwise, the multiplier effect and community benefits can not be achieved.
A lot more education is needed. We should have aligned ourselves with every community through a massive communication campaign, which has not been done. The lesson learned is that what can prevent a technology that provides great value from not having a massive impact on everyone are the problems of cultural adoption. It is a project with a very high level of innovation, which has worked in some areas and not in others due to a lack of cultural adoption or lack of time due to the health emergency, but it shows us the way to follow.
Published by OPINNO © 2022 MIT TECHNOLOGY REVIEW spanish edition