AI: A Powerful Tool

With great power comes great responsibility — so asserts Suilan Estévez Velarde, a professor of Artificial Intelligence at the University of Havana, who presented an intriguing project during the World Youth Summit of the International Telecommunication Union, which gathered about 400 delegates from all regions of the world in Varadero, Matanzas, Cuba.

"The project is called 'Autogol,' like in soccer. We named it that because it essentially does what a Machine Learning specialist would do — finding the best solution to an artificial intelligence problem where you have data and want to learn patterns from it, classify images, summarize texts, make future predictions, identify entities being discussed, and so on. These are the types of pattern recognition problems widely used today, especially in industry. This tool automates that process. Instead of training a specialist who needs ten years of experience — typically PhD students or doctoral experts, which takes a long time and significant investment — you would have a system that people with only basic programming knowledge could use to find the best solution to the problem."

What stage is the project in now?

"At the moment, we have a fully open-source library available on GitHub. It includes over 130 algorithms from around 10 different Machine Learning libraries used by specialists. We keep adding more. Currently, we are also working on eliminating the need for programming knowledge altogether — so that users can simply describe their problem, provide their data, and the tool handles the rest. It’s still a work in progress, but it’s already functional for several tasks. Of course, it’s a tool designed for developers, not the general public; you still need to program to use it — but it’s completely free and open."

Where is it being used currently?

"We’ve used it in several of our projects, particularly in industry-related tasks and research."

At the university itself?

"Yes, because one of the challenges we face when running experiments or solving a problem is finding an initial solution — or even validating multiple experiments simultaneously. This tool can test several solutions and later show you what worked. The output isn’t in plain text but in numbers and statistics — data you can use.

"Additionally, we’ve collaborated extensively with the University of Alicante and are now starting to work with the University of Oriente. It’s a tool you need to learn to use, and we’re attending events like this to promote it, ensuring more people have access to and awareness of its availability."

Have you received any feedback after presenting it at the Summit?

"Yes — two or three different people from various institutions have approached me, interested in understanding what the tool can do and how they might incorporate it into their processes."

What are the advantages and what warnings should we keep in mind when using AI?

"AI is a tool that’s here to stay, primarily because it’s incredibly powerful. With great power comes great responsibility. It allows us to do things we couldn’t achieve on our own — we can’t analyze millions of data points or comb through the entire internet to find the information we need, but an algorithm can do that for us. AI can expand our capabilities, give us what feels like superhuman abilities, and help countries gain resources and improve problem-solving capacities. However, it’s a tool we must use properly.

"I believe the greatest warning regarding AI today is that, while it can do tremendous good, it’s often misused — unfortunately. Many young people use it without understanding its limitations. For example, when using language models like ChatGPT, people interact with them without realizing they can make mistakes or 'hallucinate' — a formal term for producing false or misleading outputs. These errors can be catastrophic. If you’re unfamiliar with the topic you’re asking about, you might not notice the mistakes. So, if an expert uses AI to speed up their work, that’s effective — but if someone relies on it to learn about a subject they know nothing about, they’re likely misusing it.

"It’s essential to use AI properly — ethically, for the benefit of people, not purely for profit. While these tools can learn countless patterns, they can also absorb and replicate biases. This happens because we, unfortunately, feed them biased data. The past isn’t how we wish it were — it’s how it was, and despite our efforts to improve things, the data still reflects that reality. As a result, we have algorithms that exhibit racism or sexism, stemming from the very data we provided. For developing countries in particular, AI offers tremendous advantages — but we must ensure it’s used correctly."

Translated by Sergio A. Paneque Díaz / CubaSí Translation Staff