It was ’68. «In those days the girls who studied mathematics all chose teaching. I, with two others, chose the application one instead: it was the breaking point.” It was 1968, the year in which Luigia Carlucci Aiello graduated from the Normale di Pisa. Today she is for everyone the mother of artificial intelligence in Italy, then she was a young student looked at askance, «also because I went through a period of chauvinist isolation. At every exam there were colleagues outside the door waiting for me to fail.” And instead…
Let’s start from the beginning. Why mathematics?
«As a child I always preferred it to other subjects, and at a certain point it was a passion that could no longer be questioned».
Not even from his parents?
«My parents were in fifth grade. My father worked on the railway, my mother was a housewife. Her teacher had insisted to my grandmother that she could continue her studies, but there had been nothing to do. She dreamed of a future for me, in turn, as a teacher.”
Why then the university?
«Because at the beginning I thought that being a teacher would be better. Then, at a certain point, things got out of hand.”
He was saying about his colleagues.
«Those who had entered with me had lost their places, I was left alone. On the positive side, I made good friends and learned a lot of things, but the professors made war on me. And for the normalists outside the door there was, I’m not saying the desire that I would do badly, but the certainty that this would happen. It wasn’t pleasant.”
Instead, the degree…
“Certain! I really cared about it, but I got bored of the Normale. And since my supervisor worked at the CNR I went there: it was my luck.”
They were the legendary seventies of hi-tech.
«In universities there was practically no talk of computers. The Cnr was freed from those logics, more than it is now. I finally felt free to do research, in particular on what cybernetics then became computer science.”
Which is the ancestor of AI.
«In Pisa the environment was very stimulating, the transition was natural. Except that talking about intelligence was considered to scare people, so our department was called non-numerical information processing. Do you understand what a play on words is?”.
From Pisa the flight to Stanford…
«The CNR left a lot of autonomy, so I had the opportunity to go twice for a total of four years to where everything was starting. It was the period of flower children who had become nerds, there wasn’t just Steve Jobs: garages were constantly popping up where ideas were boiling. I ended up in the laboratory of John McCarthy, the man who coined the definition of artificial intelligence. I was already in the future.”
Another world.
«For those who came from Europe like me, it was a cultural shock. In those laboratories there was a way of working that was alien to us: they were open 24 hours a day, we had fun and there was a technological gap of at least five years. Above all there were computers, the real ones.”
An example?
«When I arrived in 1973, Arpanet already existed. The idea of being able to exchange documents with some colleagues from MIT in Boston was incredible. Of course, there was less privacy protection. But it was science fiction.”
Other memories?
«Seeing what is today Silicon Valley grow around Stanford. Large companies opened their offices there and so we moved from large computers to personal ones. And the internet arrived.”
The 80s and the return to Italy.
«And in the past: there was a lot of resistance to change. We had to wait until the end of the decade to see PCs on the desk. Meanwhile in 1981 I had won the competition at the Sapienza University of Rome and the following year I entered the Dis which today is called Diag, Department of computer, automatic and management engineering. Then in 1991, the chair of artificial intelligence.”
The beginning of everything.
«In reality we talk about AI from Alain Turing’s studies: from one point of view it has never changed. We can then begin its development with the arrival of calculators in the 1950s, but the concept remains that of Aristotle’s syllogism. It’s about wanting to mechanize human actions using common sense logic.”
Put like that it seems simple.
«In fact it isn’t. At first the studio was divided between software and robotics, and it was fun to see what those who built intelligent devices were doing. At the end of the century the two parts reunited and we began working on the autonomy of machines.”
Also thanks to his studies.
«The group I was part of spread out to teach in universities throughout Italy, even though AI was still not held in high regard. In fact we are talking about winters of artificial intelligence.”
How come?
«The lack of results weighed heavily. The theorems proved that certain things were very difficult, and we were considered to be the ones who wasted time on useless things. A former student of mine recently confessed to me that her father wouldn’t have wanted me to be his advisor because, he told her, you’ll never find a job. The father was a professor colleague of mine in the scientific field.”
Then came summer.
«The results have arrived, the speed that is changing everything. In some ways it’s impressive.”
AI is also posing ethical problems.
«Great importance is given to this, but these are problems that have always existed, even when the first ideas became software. For example: if you create an open system for medicine, who has responsibility if a patient dies? Who made the program? The doctor? Both? Ethics depends on people, who are the key to everything.”
In what sense?
«I always said to my students: let’s say an AI system convinces you to invest all your money in LCA and you trust it, who is to blame if you go bankrupt? Ah, LCA stands for Luigia Carlucci Aiello, and I’ve never had a company.”
However, people today tend to believe everything that arrives on their smartphone.
«This is a very serious thing. We must relearn to have a critical spirit, the same as when we studied from books. Did we by any chance then take everything as true? At school they taught us to read and analyze. Today more than ever we need training.”
«AI is like a hammer: it depends on how you use it»: this phrase is his. Don’t we talk too much about its dangers and less about the benefits?
«Very true. I used that metaphor to say that tools are created with objectives, but then they can also be used for other purposes. Of course AI can be dangerous. Especially if it is used by an incompetent person, or by an uncontrolled and non-transparent system.”
Especially now that she’s learning from herself.
«So what? For example: would you get into a car and drive it if I told you there had been no testing? Would you agree to take a drug that has not passed a test? Behind the reasoning machine there must always be someone who directs it.”
But can a robot ever have a conscience?
«The debate is open. It depends on what consciousness means: if it is having a sense of oneself, one enters the field of free will. There are those who are convinced so and those who think it is not possible like Federico Faggin, the inventor of microchips. I tend to think like him. And in any case I don’t blindly trust machines.”
Why?
«I believe in their power, but also in the rules. You know what they say: America invents, China copies, Europe regulates. Well, the fact that the United States is a little reluctant to set limits is a problem.”
Did that recent graduate from the Normale ever think she’d see all this?
«Yes and no. But in the end they are always algorithms. Today everything is said to be artificial intelligence, but it’s not as if it invented itself. It was our programs, which come from other programs in which errors have been corrected. So in the end yes: getting this far was predictable.”
What will happen now?
«It’s difficult to say. But I’m worried about something else.”
Which?
«There is too much discussion about the risks of AI instead of focusing on the climate. We can control what the machines do, but due to the environmental crisis we can reach a point where it no longer depends on us.”
What to do?
“I’m afraid we’re already there.”