Alicia Troncoso, president of the Spanish Association of Artificial Intelligence: “There are no machines with common sense”

Alicia Troncoso Lora, born in Carmona (Seville) 48 years ago, has been receiving recognition since the end of her university education, which she spent in Seville, New York and California. She is a professor of Computer Languages ​​and Systems at the Pablo de Olavide University, she is now president of the Spanish Association of Artificial Intelligence (AEPIA). He defends that this tool has to be understood as “a country project” that includes, in addition to the technological vision, a perspective social, humanistic, legal and ethical. “It is essential,” she says.

Ask. A basic definition of artificial intelligence (AI) refers to systems and programs with human-like capabilities. Will they replace us?

Answer. It is a fear that everyone has inside them seeing the advances in artificial intelligence. But replacing a person is complicated because a person has many qualities that, today, machines cannot have. There are no machines with common sense. Replacing us is a long way off, because the artificial intelligence that is advancing right now is specific, that is, algorithms and machines are built that learn to perform specific tasks. If the task changes, you need another machine or another algorithm, while a person, alone, is multitasking, he does everything because human intelligence is generalist. Yes, they are going to execute certain tasks that people do.

The machine must be told what is good and evil. She doesn’t know. She alone she cannot distinguish

Q. But will they be able to distinguish good from evil?

R. A machine will do what we design. The machine must be told what is good and evil. She doesn’t know. By itself it cannot distinguish, but it is in the ethics of the entire sector—researchers, developers, companies and governments—that this artificial intelligence is designed meeting certain requirements so that we do not later find ourselves involved in problems related to ethics.

Q. Which is it?

R. There are many and they are related to justice and transparency. Intelligent systems will make decisions or help make them. For example, they can be used to grant credits and it must be guaranteed that they act without bias, because we have already seen some systems that act in a discriminatory manner. Intelligent systems must guarantee equity, that there is no type of bias or discrimination. It is a great challenge related to justice. The other, and on which there is a lot of research right now, is transparency: if they are going to make decisions that affect people, we have to know how they work. If they do not grant credit, let them know why.

Q. Artificial intelligence has gone from being a discipline for experts to becoming a tool that even schoolchildren use, like ChatGPT. Is it beneficial?

R. Great benefits are beginning to be seen and much more will be seen. There are applications for neuroprostheses that allow people who have not seen for 12 years to distinguish colors or to allow the mobility of an articulated arm only with thought or to activate neurons from a memory, which is related to Alzheimer’s. In the not too distant future, they can help with many health problems.

The raw material of artificial intelligence is data and in this sense there are security and privacy risks. It also has a labor impact. That is why regulation is so important.

Q. The case of Almendralejo, where some teenagers recreated nudes of classmates with artificial intelligence, shows that there are risks.

R. There are many risks and we see them every day. The raw material of artificial intelligence is data and in this sense there are security and privacy risks. It also has a labor impact. That is why regulation is so important.

Q. Algorithms decide the price of a ticket, our social relationships, our purchases… Has artificial intelligence crept into our lives without us realizing it?

R. We are surrounded by artificial intelligence and we have it so internalized that we do not distinguish it, we are not aware of it or we do not give it importance. How many years have we used Google Maps? This application is artificial intelligence. Smart vacuum cleaners have the plans of your house. The publicity that reaches me is different from that which reaches another person. All of this has artificial intelligence and, of course, it is conditioning our consumption habits, our leisure, and many aspects of each person’s social life.

We are surrounded by artificial intelligence and we have it so internalized that we do not distinguish it, we are not aware of it or we do not give it importance

Q. Where are we going?

R. I think green artificial intelligence is now being advocated. AI also consumes energy and is also contributing to this climate change. Right now there are very powerful initiatives to make artificial intelligence green and to ensure that the energy consumption of those algorithms that all the large platforms have underneath, such as Google or Meta, is as low as possible. Also, due to the upcoming regulation, we are working on artificial intelligence designed and centered on the person, to avoid all this collateral damage. Progress is being made on productivity in the field… Artificial intelligence is like everything else: used well, it can bring us benefits. But, to use it well, you also need training in values ​​and ethics to know what is right and what is wrong.

Q. Therefore, as he said, regulation is necessary.

R. It is very necessary. Right now there is already a proposal for an artificial intelligence regulation in the EU that classifies tools according to whether their impact is high, medium or low. Based on that, there will be defined requirements for artificial intelligence tools. This general framework will advance so that it includes all the cases that may arise.

Legislation is always slow while technology is the opposite

Q. He comes late?

R. Technology advances very quickly and law and legislation are slow because laws have a lot of impact and have to go through different organizations and be subject to discussion to guarantee that they truly fulfill their purpose and do not cause harm to people. Legislation is always slow while technology is the opposite.

Q. Some managers of technology companies and intellectuals advocated stopping until they had an agreed upon framework. It is necessary?

R. That petition, signed by many executives of large companies dedicated to artificial intelligence, arose when ChatGPT exploded. There are many cross interests in these companies. When Google or Meta brought an artificial intelligence product to the market, they did not worry about the damage that artificial intelligence could do. They were worried when ChatGPT emerged because it was released by another company that was not theirs. There are many conflicting interests in that request and it must be quarantined a bit. Artificial intelligence must not stop, but it must advance in accordance with regulations, guidelines, and ethics, to ensure that it is centered on people and is used for their benefit.

You can follow EL PAÍS Technology in Facebook and x or sign up here to receive our weekly newsletter.

_