🤖 Exclusive Interview with Enric Trillas 🎯 “AI models reason, but they don't think”

🧠 “I’m recognized as the pioneer in Spain of the fuzzy logic system, methods for dealing with language imprecision”

📐 “In 1998 I introduced the first mathematical formalization of conjectures in the field of AI — I dedicated myself to it deeply”

Enric Trillas Ruiz (Barcelona, 1940) is the father of fuzzy logic in Spain — a theory that allows working with imprecise or gradual values such as “hot,” “cold,” “a bit,” or “very,” and is fundamental in AI systems that must interpret data with uncertainty or ambiguity. A PhD in Mathematics and professor at both the Polytechnic University of Catalonia and the Polytechnic University of Madrid, he was the first president of the Spanish National Research Council (CSIC) under Felipe González’s governments (1984–1988). For the past 19 years he has lived in Asturias, is professor emeritus at the University of Oviedo, and has been retired since 2017. Claudi Alsina is considered his most outstanding disciple and supervised the thesis of Ramon López de Mántaras at UPC. A modernizer of CSIC, during his presidency the Center for Advanced Studies in Blanes was created, and in 1985 he entrusted López de Mántaras with founding the AI research group in Blanes — the seed of today’s IIIA. Inexplicably, this Catalan scientist has not been awarded the Creu de Sant Jordi, despite being one of the founding figures of AI in Spain.

What did you teach?
Since I moved to the Polytechnic University of Madrid, I became a professor of fuzzy logic, meaning methods to handle the imprecision of language. We speak using many imprecise words: tall, handsome, pretty, far, nearby...

Is it related to computer vision?
There are applications, yes. Fuzzy logic is applied in many places. Essentially, in control systems. For example, the control of a bridge crane, a train, an inverted pendulum... The fuzzy system does it perfectly.

Are you a pioneer of fuzzy logic?
In Spain, yes. They recognize it.

Who is the father of it worldwide?
Professor Lotfi A. Zadeh, of Iranian father, Russian mother, born in Azerbaijan, raised in Tehran, fled to the United States in 1944.

Is he from Berkeley?
Yes.

Did you meet him? How did it go?
Yes, it’s complicated. One of my disciples, probably one of the best, Claudi Alsina [father of former Minister of Foreign Affairs Victòria Alsina], thanks to my friendship with Austrian professor Karl Menger, who fled to the U.S. before the war when Hitler invaded Austria, went to do a postdoc at the University of Massachusetts. After six months, I went to visit him, in 1976. Then I took the opportunity to go to San Francisco to meet Lotfi A. Zadeh, but he wasn’t there. No one knew where the hell he was. He traveled a lot. In 2012, five years before he died, he traveled 250,000 miles. The point is we didn’t know where he was, and I left a message that I had been there. In 1977, I was organizing in Barcelona the first World Conference on Mathematics in the Service of Man, with Antoni Ballester. I invited Zadeh, and he came! So, I met him in Barcelona.

Did Ramon López de Mántaras meet Zadeh then?
No, Ramon met him in Toulouse. There’s a lovely Catalan there, now retired, Joseph Aguilar-Martin, whom we all knew. We used to go there a lot. I had the visit of a very valuable Japanese man, Michio Sugeno, who died about a year ago. Everything done in control theory uses his ideas. Michio Sugeno came here. Why? I went to Toulouse to meet Josep Aguilar and there I met a young Catalan guy running around who had studied technical engineering in Mondragón — that was Ramon. The thing is Sugeno comes to Barcelona, and I take him to Toulouse. And, boom, there’s Ramon again. Meanwhile, in Barcelona, I was helping Gabriel Ferraté [1932–2024] to create the School of Informatics. None existed back then! And when Gabriel Ferraté became General Director of Universities in Madrid, he got obsessed with the idea. Ramon Puigjaner, myself, and three or four others helped him. I typed the decree of creation [of the informatics schools] at 2 a.m. on a Wednesday at the Polytechnic’s rectorate. He took it and the next day it went to the Council of Ministers. That’s how it happened. From my typewriter to the Council of Ministers.

Do you have the Creu de Sant Jordi?
No. I have the Monturiol Medal.

And what is more important?
The Creu de Sant Jordi.

Why haven’t they awarded it to you?
I don’t know. I’d be very happy to receive it, but I’m not upset. I have many medals, a ton of them. I’m even a Grande Ufficiale of the Ordine al Merito della Repubblica Italiana.

How did that come about?
When I was president of the CSIC, I did everything I could to improve relations with the Italian CNR [Consiglio Nazionale delle Ricerche]. Both institutions are very similar, from the same era. That one was founded by Mussolini, and ours by Franco. Many years have passed, and everything is very different. Relations weren’t great. I got obsessed and traveled to Italy many times. I was received by the then-president of the CNR, Professor Luigi Rossi Bernardi. I was wearing crepe shoes and apparently built up static electricity. I walked down a huge hallway with carpet. He came out of his office to greet me. We shook hands, and I gave him an electric shock that made him jump. He literally jumped. I thought, “We’re screwed. No chemistry here.” But there was electricity! We became friends, very good friends. From then on, we did many things together. For example, when we founded the Center for Advanced Studies in Blanes, I imposed a condition that everyone accepted: apart from the Spaniards, like Ramon, there would be two foreigners. I didn’t trust them because they were all very young. Ramon was a kid! I was afraid they wouldn’t take it seriously. I thought that if there were two foreigners, they’d take it more seriously. I got the CNRS to send Joseph Aguilar, who had influence over Ramon, and the Italians sent a friend of mine, Settimo Termini. They both stayed the first year.

What are you most proud of from your four years at CSIC?
Of throwing the lottery drums out the window.

What? What lottery drums?
The ones used in hiring. It was a huge job to form the committees. You can’t imagine what it’s like to organize a hundred committees. With the drums, all were from CSIC. Pure inbreeding. Only people already inside got in. No one from outside stood a chance. Ramon would never have made it under that system.

How many people were at CSIC?
At that time, there were 110 institutes and 3,000 people.

You created the Center for Advanced Studies in Blanes?
Yes, just like the National Center for Microelectronics, the Center for Economic Research, we also started helping Barcelona City Council with the Botanical Garden. We did many things.

After the drums, I suppose the impact was mid-term...
Very big. The impact of the new selection system for researchers was huge.

Was it democratizing?
It was opening up. Fresh air. The worst enemy of universities or any research center is inbreeding. Inbreeding leads to mediocrity, and mediocre people only want to surround themselves with mediocres. It gets worse and worse.

Thanks to you, CSIC raised its international standards?
It has changed a lot, and I haven’t been president for many years. There have been many since then. It’s not all my merit, but I take credit for starting it.

You were appointed by Minister José María Maravall. How did they think of you?
Through [Alfredo Pérez] Rubalcaba. We were friends. [Laughs]

How did you know each other?
From Valldoreix.

Were you a member of the PSC?
Yes, I’ve always been a social democrat. In the PSC, I was responsible for science policy. I organized a meeting of scientists in Valldoreix to talk about the state of science in Spain. It must have been 1980 or 1981. Many people from Barcelona came. From Madrid came Carmina Virgili [Barcelona, 1927–2014], then a professor at Complutense, and Rubalcaba, a very young chemistry professor. They came, and we became friends!

There was chemistry there.
Immediately. With Rubalcaba, instantly.

So Rubalcaba proposed you?
Yes, he and Carmina Virgili. When they won the elections [October 28, 1982], Carmina became Secretary of State for Universities and Research, and Rubalcaba became her chief of staff. And they proposed me.

Did that derail your research career?
No. It’s more complicated. When they won the elections in ’82, I was appointed Spain’s delegate to the OECD’s Committee for Scientific and Technological Policy in Paris. I had to go every week, but I taught classes on Monday and Friday mornings in Barcelona. I kept that up from January 1983 to May 1984, when I was appointed president of CSIC. At CSIC, yes, I had a research break. The first six months I couldn’t do anything. I didn’t know CSIC. I knew Antoni Ballester [1920–2017], from fisheries research in Barcelona. We were very close friends. By the way, Antoni was the man who took us to Antarctica.

Oh, really?
When I was appointed president, since I didn’t know CSIC at all, I wanted to bring someone I trusted who knew it. I thought of Antoni, I didn’t know anyone else. And he said: “I’ll come, I’ll help you, but you have to help me. I want to go to Antarctica.” And we did. We made three trips to Antarctica. We founded the Juan Carlos base back then.

Any other major accomplishments?
Microelectronics, biotechnology, economic analysis...

What was Spain like in the early 1980s in terms of science?
Sad. In the 60s and 70s, in terms of scientific output, Spain ranked 29th in the world. Now we’re between 9th and 10th. Think of who’s above us: the United States, the UK, France, Germany, China, Japan... That’s not bad. In fuzzy logic, my specialty, at the beginning of this century, in 2005, we were the third country in the world.

How do you explain that?
By working like beasts. There’s no other explanation.

Were they all your disciples?
Not all of them.

The fruit of your seed?
Even the center in Granada, which is very good, was born because of me. Many people from Granada came to the courses I organized in Barcelona.

After fuzzy logic, did you do other important individual research?
Yes, but whether it was important, I don’t know.

Please explain.
Much of my work until I left Madrid was with Claudi Alsina. We published over 70 articles together, with very mathematical results, and I think some of them were very good. When I left Madrid, Claudi was already fed up, and also, he joined the Catalan Government with [Andreu] Mas-Colell. He told me: “I can’t do this anymore.” What can you do. I had to decide whether to continue with that, which we had developed a lot, or explore something I had been thinking about — the study of reasoning and language. Since 2006 I’ve mainly focused on that. In 1998, I introduced into the AI world the concept of the conjecture. It’s the first mathematical formalization of conjectures that exists. I discussed it a lot with Settimo Termini, and that’s what I’ve dedicated myself to.

Is that what’s now called reasoning in natural language models?
Yes, yes.

Do the models reason or not?
Yes! What they don’t do is think.

What’s the difference?
Thinking is automatic. You and I will think until we die. Until our brain stops.

And reasoning?
Reasoning is regulated. It requires two things that thinking doesn’t: your will — you must want to reason. “I want to find this, I want to prove that.” For thinking, that’s not necessary. And then you need rules. You don’t reason just any way. You must preserve truth.

Is reasoning more secure and reliable than thinking?
Yes, but reasoning is a phase of thinking. It’s not different — it’s a specialization of thought to tackle problems. Thought can be completely free or directed. You can direct your thinking: “I want to go interview Trillas.” Another thing is when you want to reason to find the structure of proteins, for example. You have to do experiments, many things.

Do you think AI might end up being creative like humans?
In the long term, I don’t know. Maybe. We’re not as spiritual as we think, mind you. Let’s not confuse things. I’m a Christian, but not Catholic. Christ didn’t say anything about this. Maybe AI will be [creative], but not tomorrow. We’re very far from a machine being indistinguishable from a human being.

This so-called artificial general intelligence (AGI) — are we far from it, or is it just marketing?
These are companies selling a product.

Do you agree with Ramon López de Mántaras, that we’re being deceived?
Yes. But we differ in one thing. He believes [machines] don’t reason.

And you do?
I believe they do. That’s the big difference between him and me.

Why are you convinced that they do reason?
His reasons — I don’t know them, because they don’t exist. Reasoning is rule-based. Machines follow rules. Thought is not rule-based. Machines cannot think. Thinking is not always digital. It is often analog. Careful! You can imitate analog systems with digital ones, but they are not analog. There are many things that make me think machines do reason, but they don’t think. The big leap in AI happened when they truly started using neural networks. That was the breakthrough.

They don’t think, but they do reason — could they develop feelings or emotions?
If the emotions are simulatable, yes. But they’ll have to simulate them somehow. How does emotion explode inside you for a woman? It just does. Will a machine be capable of that? Of seeing a girl and feeling something? Of talking to that girl and feeling more things? Of cohabiting with that girl?

More and more people use AI as psychologists or personal assistants and get emotionally attached.
I don’t know what to think. I don’t know what they’re looking for. As support, these systems find a lot of information and synthesize it. That can be very useful! But you must use it yourself. The center I went to in Asturias was shut down. Sergio Guadarrama, who works at Google, one of my best disciples, played a joke on me by sending me a report about the closure of the center. I read it and it was quite good. Not perfect, but pretty good. And he tells me: “It was made by Gemini [Google’s AI].” If I hadn’t known anything about that topic, it would have been useful. What I shouldn’t have done is believe it entirely. But as a guide — if it’s for a patient, for a prescription — it can be useful.

You’re 85. Do you use these artificial intelligences?
No. I’m old-school. I write with a pencil.

When did you stop?
With neural networks.

Are you afraid of the path AI is taking?
Yes, it can be scary. For example: deceptive photos. Fake news. It can be terrible. In Spain, it could provoke a civil war.

How can it be prevented?
By control. Who has to do that?

Is the European law enough?
Commercial interests are huge. Very, very huge.

Is democracy taking a back seat?
I’m afraid so. Without democracy, there are things that, in my view, are impossible. The freedom to create.

In 2025, are we at that point in Spain?
There is a persecution against this government! Acoso y derribo.