Valter Fraccaro: "AI without ethics is not true intelligence"

Valter Fraccaro: "AI without ethics is not true intelligence"

Reconnaissance on the state of the art and the future of Artificial Intelligence in the world with the expert manager from Veneto at the SAIHub summit in Siena

Valter Fraccaro, President of SAIHub, acronym of Siena Artificial Intelligence Hub, is often the guest of public events on Artificial Intelligence
Valter Fraccaro, President of SAIHub, acronym of Siena Artificial Intelligence Hub, is often the guest of public events on Artificial Intelligence

Was born in Torino, his family roots go back to the plateau of Asiago, but chose the Toscana to live and work; is considered one of world's leading experts in Artificial Intelligencebut of itself Walter Fraccaro he prefers to tell what really matters to him: “I'm someone who rides a bike, who also reads the cheese menu, who thinks that if you didn't learn anything that day you did wrong to get up, that if we have so many verbs like studying, learning, knowing, understanding and explaining it's because everyone it is based on the previous one, who is 56 years old and who appreciates young people more than those of his age…”.
President di SAIHub, Short for Siena Artificial Intelligence Hub, pole of global importance in the field of Life Sciences, is convinced that the Life Sciences are essential for achieving the Sustainable Development Goals defined byONU, and therefore "I work to promote the increasingly widespread use of Artificial Intelligence in companies and to organize orientation activities for young people, so that they can become passionate about science and choose it as a profession".

As a consultant, he has always lent a hand to companies in the definition of Artificial Intelligence projects aimed at creating a real competitive advantage, taking care of the technological, ethical and social aspects, with particular regard to Sustainability. He also acts as an advisor for private companies, foundations and business consultancy firms in the field ofArtificial intelligence applied to R & D, the so-called Research and development. In this context, he has written six projects on the subject of Life Sciences, each of which has obtained public and private contributions and funding.
In his professional trajectory, he has crossed paths with various Italian companies and localities, always in positions of great responsibility, some still current, others relating to the past, all in a period of time coinciding with the last two decades: Monte dei Paschi Foundation of Siena, bba Delivering The Biotech Promise, Parolin Legal, Achilles Vaccines, Life Foundation, 4Sciences, Dedagroup ICT Network and the prestigious University of Padua.

The most famous definition of Artificial Intelligence, in Italian, is that of the engineer Marco Somalvico, one of the pioneers of the matter. He says: "It is a discipline belonging to computer science that studies the theoretical foundations, methodologies and techniques that allow the design of hardware systems and software program systems capable of providing the electronic computer with performances that, to a common observer, would seem to be of exclusive pertinence of human intelligence”. Is it a phrase that is still current or would it be worthy of a "coupon"? And what is the "explanation" that an expert like Valter Fraccaro gives of AI?
“It's a definition that, with that 'they seem', hits the mark. We are used to attributing intelligence to those who can solve complex problems and, therefore, when we see a computer beat a chess champion, we think it is exercising some form of intelligence, and this is where the misunderstanding arises. Although it cannot be excluded that in the future it will be possible to transfer the biological capacity of the brain to produce intelligence to a machine, today we are very far from something similar. The great potential of AI is to be able to correlate many different data with each other, something impossible for human capabilities, but typical of a computer. In short, we humans have qualitative mental abilities (we perceive the pleasure that Beethoven's Ninth can give us), the calculator is much better than us on a quantitative level (it can analyze hundreds or even thousands of factors and, for example, show us among the 10 to the sixtieth possible molecules those 10 on which to concentrate efforts in order to cure a certain disease). As for the definition of Artificial Intelligence, I think it is better to adopt a standard one, so as not to run into difficulties when working together. The one given by the European Union may be a bit long-winded, but it guarantees to last over the years. As for me, I often use the expression 'mathematical-digital tools' to summarize all the tools we use to create AI solutions”.

Who is Valter Fraccaro? And how would you describe yourself in general as well as in relation to his activity in the field of development of "Artificial Intelligence"?
“I find it much easier to describe AI! Let's say that my entire professional career, now almost forty years old, has taken place in Information Technology. I started as an apprentice, first became a private manager and then a public one, I founded two companies in the sector. Most of all, I've been very lucky to work with great masters, people who have pushed me to exercise a certain critical sense and to solve problems by grasping their crucial point. In this sense, and to mention just one name, I quote Professor Gianfranco Bilardi, with whom I had the pleasure of working daily during the three years I was director of the University of Padua: by example, he taught to use logic in a profound and acute way. Whether I've learned enough remains to be seen! He is one of those people who does not like the spotlight, but in the field of theoretical computer science, the one that pushes technology forward and requires maximum logical commitment, he is known all over the world, having opened doors to thinking that seemed to resist every effort of the human mind. As for Artificial Intelligence, I got there starting from my profession, ICT, and from my passion, science. The application of Artificial Intelligence to scientific research is the leading result obtained in the past several decades, since that expression was invented, when it meant something different from today. Today I combine these experiences with the many years spent in designing and providing IT solutions to companies of all types, sizes and sectors, so as to make the most of AI, which is a tool to improve businesses, which remains the goal . And today, as never before, improving businesses means improving the world: if we want Sustainable Development, we must make this simple concept our own".

What can you tell us about the SAIHUB “Siena Artificial Intelligence Hub”? What is its institutional mission, as well as that of the partners and founding members, and how do they all fit into the growth trajectory of AI?
“In Siena there is an economic district linked to the Life Sciences that has no equal in Europe in terms of concentration of knowledge and skills in the research and development of new drugs, medical devices, methods and technologies for the agri-food sector. Also there is a university that has levels of excellence in the same areas and in Artificial Intelligence. It is from this basis that a couple of years ago the Monte dei Paschi di Siena Foundation began to think about setting up a center dedicated to AI to encourage the introduction of these methodologies in the Life Sciences sector, so as to strengthen the economy attracting people, companies and capital to the area capable of guaranteeing the position of international pre-eminence of the Sienese and Tuscan ecosystem in the future as well. Once the idea was proposed by FMPS, the University of Siena, the Toscana Life Sciences Foundation, the Municipality of Siena and Confindustria Toscana Sud immediately joined. At that point and with the same founding role, the SAIHub Business Network was set up, to which around thirty companies from all over Italy and beyond belong today: the Ticino-based has also recently joined which, with an Italian , created Tecnoscientia, which is based at SAIHub. We are dedicated to various activities aimed at young people and businesses. This summer we created the first Italian Summer School on Artificial Intelligence and Life Science: ninety girls and boys who had just completed the fourth year of high school were able to participate in five days of orientation on these subjects for free. Our intention was first of all to create social awareness and, secondly, to make young people passionate about science, so that perhaps they may enroll at the University of Siena and become the researchers of tomorrow. We will propose this initiative again next summer and we expect at least 200 boys and girls. Furthermore, we have funded several hundred thousand Euros of scholarships for the best university talents and, finally, we have created economic prizes (up to 10.000 Euros each) for those who, having obtained a master's degree or doctorate, enter a be part of the teams of the companies of the 'network' and remain there for at least one year. This latest initiative aims to make these young people understand that in Siena there are opportunities for developing their scientific and career knowledge, which are not inferior to those found abroad. Thus they can choose to go outside Italy, but they know that they are not forced to and, in any case, that when they want to return, they will be able to find a land ready to let their abilities flourish. We then began to financially assist the AI ​​projects presented by the various companies in the 'network' in collaboration with each other. Seven of these projects are shared with the University of Siena and have earned the regional funding of several research grants.

And Italy gives the green light to the plan for artificial intelligence

Valter Fraccaro is President of SAIHub, an acronym for Siena Artificial Intelligence Hub, a hub of global importance in the field of Life Sciences
Valter Fraccaro is President of SAIHub, an acronym for Siena Artificial Intelligence Hub, a hub of global importance in the field of Life Sciences

What is the "highest point" currently reached by Artificial Intelligence, the goal beyond which it has not been able to go so far? The question is applicable to the whole world and to Italy…
“The answer can only be subjective, in this case. In the meantime, we could start from the highest values: saving human lives, for example. In this case, it is right to think about how relevant AI has been in the research and development of vaccines against COVID and drugs to combat them. Creating effective vaccines in less than a year would have been impossible without the contribution of Artificial Intelligence. As Rino Rappuoli, an extraordinary Italian scientist and reference in world vacciniology, said, if ever a next virus threatens humanity as the Coronavirus is doing, thanks to AI we will be able to identify the characteristics of the necessary vaccine within the 24 hours following the moment in which we will know its DNA. Still remaining in the field of health, let's try to think about what 'personalized medicine' means. Today we all use the same medicines, men and women, young and old, wherever in the world we live (at least where medicines are available…) and regardless of our habits, diet, environment. So we happen to go to the doctor, he orders us to take two tablets for seven days and then we go to the pharmacy and buy a pack that contains 30. Pure waste to produce, package, distribute, dispose of the excess. Personalized medicine, available within a few years thanks to AI, will cure us better because the drug will be made for the individual patient, in the measure and form that best suits him. Treating in this way will give better results, which means less chances of ending up in hospital and, if you need to go there, staying there less time, thus reducing the need for huge buildings to be heated in winter and cooled in summer, with appalling maintenance costs, pollution , land use, etc, etc. Healthcare costs are worth 10 per cent of world GDP, therefore every saving is worth billions: with AI we can save a lot, as well as take better care of many more people. I say 'AI' for simplicity, but let's always remember that it is a means: humans capture the results, applying ingenuity and conscience. Neither Artificial Intelligence nor man works miracles, but if one uses the other well, then great goals can be achieved. Today the entire world population, in its economic, social and political components, is engaged in the 'transition', i.e. the path to reach a system which creates, through sustainability, a condition in which human life is better without this mean worsening that of the entire planet, as we have done for the last century and a half. AI is a tool that can help a lot on this journey”.

What do you think of the Strategic Program for Artificial Intelligence 2022-2024, the result of the joint work of the Ministry of University and Research, the Ministry of Economic Development and the Ministry of Technological Innovation and the Italian Digital Transition, launched on seen from Palazzo Chigi? Are you convinced of all twenty-four programmatic points communicated to the public, of the relationship between means and objectives, or would you have made adjustments to any of them?
“First of all it must be said that writing that Program must not have been easy at all and one must take this into account when criticizing it, because the work of others is often underestimated, thinking that it would have been easier to do it better. Having said that, I believe that that document, being of direction, goes in the right direction. It will then be the legislator's task to ensure that its application really favors the development of companies that create AI solutions and those that will use them. Compared to the Program, I think it would have been even better if it had included the international debate on ethical issues regarding Artificial Intelligence. I am convinced, after years of work and continuous dialogue with many operators, experts and companies, that the real success of any AI project is closely linked to the depth of the ethical analysis of the context in which it is located and its possible outcomes".

A quality forum for AI and business

Umberto Eco and Valter Fraccaro in a public event in 1989
Umberto Eco and Valter Fraccaro in a public event in 1989

As things stand, how much “intelligence” and how much “artifice” is there really in AI? Is man still at the same time the limit and the engine of the attempt to reproduce the synapses of the brain, the measure of everything?
“Short answer: 'if it's intelligent, it's not artificial; 'if it's artificial, it's not intelligent', as Luciano Floridi says. Longer answer: 'if we consciously talk about intelligence, then we need a precise definition of at least human intelligence, and we still don't have it'. It is difficult to determine what intelligence is except in relation to consciousness, and so the matter becomes even more complicated. Then we should get in the way of desires, which create goals. We can say one thing: to date, machines have no conscience, no desires, no goals, therefore no intelligence. The 'artifice' remains, yes. Dante Alighieri used this word to say something of great value in the materials and ability of its builder, complete in his observability. Who knows if he would like to see how we have shaped mathematics and electrons to solve problems that are too large for our minds. AI remains a decidedly human phenomenon and I fear that giving it so much power that it can scare us is a way of hiding our responsibilities. Like any technology invented by man, even if it was the first stick with which one of our ancient predecessors made an animal of the savannah escape, we are responsible for its use. And the better we use it, the more we become 'sapiens'. As for the synapse, it seems to me that the twentieth century dream of using electronics to simulate the brain is no longer the goal of Artificial Intelligence, at least not in its practical and industrial form. It was called 'cybernetics', remember? It was like having toured our cultural history: for the Greeks Kybernetes was the one who steered the ship and, by extension, also the city... now we write cybernauta, with that y that seems English to us, and instead it is the fingerprint of the our ancestors. We are always steering the ship but sometimes we fear the ship itself more than the sea. AI is only the newest model of boat: in front of it is the ocean of things we don't know and of our conscience. For that we need ethics, the compass to know where we are and decide where to go. We are the 'measure of all' not because we are better than other creatures, but because we have no other choice. We can only look at the universe with all our limits and, it can be said, we are the species that has most known how to use technology to go beyond them”.

What is machine learning (also called "machine learning" in English) and why is it a branch of Artificial Intelligence that is gaining increasing importance? Did programming move on to training, with the result of making robotics the true field of application of these technologies?
“Let's start at the end. Robot comes from the Czech 'Robota', which means 'hard work' and, before that, it meant 'servitude, slavery'. When I hold my seminars in companies, I always ask those who listen to me what work that technology has allowed us to entrust to machines they would like to go back to doing by hand. Nobody, after all. Because we, we humans, don't like work understood as forced effort. In the Bible, work is the punishment that God inflicts on Adam and in the different languages ​​the word that means 'work' always has a root that speaks of slavery. In the end, 'machine learning' is the way we have devised to transfer what we call 'experience' to machines. It is enough for a child to observe three cats to call the next one he sees a 'cat'; a computer needs to show thousands and thousands of photos of cats before it recognizes one and identifies it as a 'cat'. Only that thousands or millions of photos, the computer examines them in a few seconds and is capable of doing the same with enormous quantities of numerical data or with books and scientific texts: it is not useful for distinguishing cats, but we are interested in the fact that can help a researcher because he was able to read in an instant a quantity of information that humans could not even browse through living hundreds of years. In summary, we make sure that the hard work, the one we don't like and we aren't even very good at, is entrusted to a machine and we use 'machine learning' (and similar techniques) to let it gain the necessary experience to get it right. He works, we think”.

Will quantum supercomputers and quantum physics give a decisive turning point to the theoretical and practical growth of Artificial Intelligence or is there another way to go? And why?
“Quantum supercomputers have completely unique possibilities and are and will be used to solve another order of problems, only rarely similar to those we face with AI. They will probably be useful one day to simulate the brain, which has a biological and electrical functioning that is certainly more similar to that of a quantum computer than to that of a conventional one, but the AI ​​we think of today and for the near future is linked to forms of ordinary computing that we use in our smartphones or personal computers. And even so, we still have a long way to go”.

Swiss competence network for artificial intelligence

Valter Fraccaro, President of SAIHub, acronym of Siena Artificial Intelligence Hub, is often the host of public events on planetary sustainability
Valter Fraccaro, President of SAIHub, acronym of Siena Artificial Intelligence Hub, is often the host of public events on planetary sustainability

Should the implementation of AI or AI, whatever you prefer, really be subject to ethical and moral limits or is the possible competition of machines and humans a false question? Will homo sapiens find in Artificial Intelligence the alter ego capable of rebalancing the accounts with nature?
“Good questions, especially if asked in succession! The 'possible competition from machines to humans' is not the focus of what we call AI Ethics. Specifically, ethics, i.e. the rules around which we agree to obtain benefits and not damages from our individual and social behavior, we need to be able to analyze the context in which we apply Artificial Intelligence, so as to be able to predict the final results and avoid the negative ones. AI has many ethical aspects that need to be managed. One, the best known but also the most discussed and for which we already have adequate means and behaviors at our disposal, is privacy. In summary, it must not be possible to consult the data in order to distinguish those that describe a specific person, except under particular conditions, controlled and accepted by the individual. However, there are more subtle aspects. We must, for example, prevent a computer from acting on data using the faulty ways of humans, such as the tendency to use bias in one's choices. Furthermore, we must avoid that the application of Artificial Intelligence, which enjoys its own autonomy in finding ways to give answers to the problems posed, does not clearly expose the logical path that guided it and led to a specific result. One can go on, but the point remains: the more serious the damage that the machine can do in automatically applying a result of its analysis, the more important it is that there is a person who decides on its execution. We can ask the machine to explain 'how' it reached a certain decision, but the responsibility for its application must always be traceable to a human. Can this condition always be created? No. Therefore, we must be willing to consider machine errors as human responsibilities. That's not big news: isn't that what we do when a factory device injures a worker, for example? We are frightened by the fact that we are not afraid of the will of a press, while we also attribute that to a machine that we believe has 'intelligence': it makes no sense, and we have explained it up to here, but it comes natural to us to think so because in our experience intelligence and will coincide. Homo sapiens, alter ego, nature… Maurizio Ferraris would like to answer here! I limit myself to saying that having become globally aware that our way of creating progress, which has led us to ensure that the last three or four human generations live in far better conditions than the thousands that preceded them, is incompatible with safeguarding the planet and the future of our descendants is the first great event of this century. We need to focus on our scientific and, by derivation, technological capabilities to change this course. We know that we have the tools to improve human living conditions without further sacrificing those of other species and of the Earth itself, and that some still immature technologies will soon give us further possibilities. On this basis, through the UN SDGs, we have made a great pact between over 190 countries, and it has never happened before.

By 2030, China will be the first player in the field of Artificial Intelligence, surpassing Europe and the United States of America: what will be the impact of this transformation on geopolitics and the world economy? Will the "hunt" for maximum environmental sustainability and the fight against global warming, the identification of new energy sources, pass through these studies?
“The right question, which starts from quantitative terms. We always closely link these analyzes to when the various countries (or 'empires', as a Venetian philosopher who is much contested today calls them, but who makes a correct analysis on this issue) are investing in Artificial Intelligence. It seems as if we were in the time of the gunboats, where whoever had the biggest ones ruled the world. I'm not sure that's how things work today, and even less that this view holds for years to come. It seems to me that skills and the ability to put them to good use are not simply a quantitative issue. I work a lot in Siena, a center of 50.000 inhabitants that is difficult to reach even by train, yet there is a cultural, entrepreneurial and scientific fabric there, so that small medieval city remains one of the world centers of research in the Life Sciences. In how many other places in the world were there more favorable conditions? Very many, but Siena is there and is working to stay on top. Can something similar be recreated by placing billions of Euros, Dollars or Yuan on the table? Maybe yes, but maybe it's better to take a fraction of that money and come to Siena, instead of hoping to recreate it in a few years. AI is a disruptive tool with great possibilities: this is why we talk about it so much. It was the same for electricity, for the train, for the Internet: some countries have invested more than others, but it doesn't seem to me that today, decades or centuries after the advent of those technologies, the others don't have them. I understand for myself that it is a reading that has a certain degree of naivety, but it seems even more naive to me, since we don't have them, to say 'we should invest trillion billion, like China does' or, worse, to say that then we should do it part, transform Italy into a holiday factory and become, as Marco Cattaneo, director of 'Le Scienze' wrote in 2012, a 'country of waiters'. As for the last questions, my opinion is that precisely because of what we have said up to now, Italy (and also Switzerland, for analogous reasons) has every possibility of having a considerable place in the world's future. Not as a political power, but as a cultural influence. The sooner sustainability will take hold, the more strength will acquire the knowledge and the ability to create new ways to use it, especially in the Life Sciences which remain the scientific-economic sector that can most easily help us achieve the SDGs, or Sustainable Development Goals, of the United Nations Organization…”.

The future is already here: always to be understood, never to be wasted…

Valter Fraccaro is President of SAIHub, an acronym for Siena Artificial Intelligence Hub, a hub of global importance in the field of Life Sciences
Valter Fraccaro is President of SAIHub, an acronym for Siena Artificial Intelligence Hub, a hub of global importance in the field of Life Sciences