Mariel Estévez: “It’s vital to understand what AI is doing, rather than delegating all responsibility to the computer.”

What led you to a career in research?

For as long as I can remember, I’ve been fascinated by how the world works. As a child, my dream was to work for NASA. I was heavily influenced by my older sister, who used to bring me astronomy materials and talk to me about mathematics. The science fiction books we had at home, like Ray Bradbury or Jules Verne, also played a huge part. As I grew older, I became convinced that a career in science was the only path for me.

In which areas have you focused your research?

In artificial intelligence, specifically within a branch known as interpretability and explainability. The goal is to understand how these systems actually work. Today, they operate as very closed 'black boxes'; the computer processes something and gives you an output, but not even the people who built the systems truly know why or what the logic behind it is. More specifically, I work on trying to predict whether a person has depression by analyzing their voice. It is a highly sensitive field. If a system is going to provide a diagnosis, you absolutely have to understand the 'why' behind it and have a clear grasp of its potential margin of error.

Where did you complete your studies?

In Buenos Aires, at the UBA (University of Buenos Aires), where I did my undergraduate degree in Physics and a postdoc in Artificial Intelligence. My PhD, however, was completed at the University of San Martín (UNSAM). I’d really like to highlight the importance of UNSAM and other newer universities located in the conurbano (the Buenos Aires suburbs), especially right now, given the current climate of heavy criticism and defunding aimed at these institutions in Argentina.

Is there a project that stands out as particularly meaningful to you?

Every single thing I’ve done since I started has contributed a small grain of sand. That said, I am incredibly proud of the project we are developing through the Marie Curie fellowship. I feel it operates on a completely different level of relevance.

How do you see the future of your field? What are the key trends or challenges?

That’s a very difficult question because everything is advancing at an exponential rate. It’s moving so fast that even software engineers at tech giants are struggling to keep pace. In recent months, AI has even begun programming itself—meaning these systems are now writing their own code. The entire paradigm can completely shift in just a couple of months.

Because of this, I believe it’s absolutely vital to understand what the technology is doing, rather than delegating all responsibility to the computer. We must closely analyze both the results it yields and the path it is taking to avoid stepping into dangerous territory.

What has receiving European backing meant for your work?

It represents a massive academic and professional leap. It allows me to conduct research in a place equipped with everything I need to thrive—professionally, alongside brilliant colleagues, and materially, thanks to the institute's computing clusters. On a personal level, it’s helping me map out my future path and define the exact areas I want to pursue moving forward.

What do you enjoy most about being a researcher?

The fact that it allows me to think, question things, propose ideas, and test them out. I love being in a state of constant experimentation.

What advice would you give to someone looking to get into research?

It’s a tough path, and thankless at times, but it has everything to do with passion and the sheer amount of work you put into it. From the moment you step into university until you establish yourself as a researcher, the most crucial thing is simply showing up, working hard, and putting in the hours.