Albert Cañigueral: "It's not a utopia for Catalonia to be able to position itself in the field of AI."

BarcelonaAlbert Cañigueral (Mataró, 1977) is a digital explorer affiliated with the Barcelona Supercomputing Center and specializes in the social and economic impact of digital innovations and access to culture.

Currently, AI is controlled by large corporations, and there are countries, like the United States and China, that are far ahead of us.

— Much progress has been made without much reflection. It's all about moving forward and running, and such acceleration is unnecessary. At the Barcelona Supercomputing Center, with the Catalan and national governments, we're working on public AI. In the same way that we provide roads, electricity, and water supplies for industrial society, we want to do the same with AI. Public AI must guarantee data traceability and transparency, because right now, companies aren't responsible for what they're doing. They take data and violate their own terms and conditions. Meta pirated 7 million books, and in its internal documents, it claims they have no economic value. It's quite unacceptable.

Is AI ever neutral?

— AI is not a neutral tool and must be carefully monitored. A lot of data goes in and out. There can be bias, and depending on how it's configured, it can amplify certain data. Control mechanisms are often put in place based on values, language, and desired results. In English, they're called guardrails. They could be translated as safety barriers.

What will public AI be like in Catalonia?

— We call it Aina, and it's a public artificial intelligence infrastructure. It's not a utopia that Catalonia can position itself in the field of AI, but it is a long-term project. It's not a technology we can deploy quickly, and we must have time to orient it. It's a fallacy that technological innovation equates to social progress. For this to happen, there must be political decisions and regulations. And they must be decided collectively.

To what extent can culture help create this AI?

— AI is fueled by data. Therefore, if we want it to have different cultural contexts and references, this data must be provided, and the Catalan language must also be taken into account. We are facing a new technological capability, and we must ask ourselves many questions. Culture can and should contribute to asking new questions, to creating a richer and more diverse AI, because otherwise we run the risk of being colonized. Furthermore, culture can help us think of different uses for AI.

And what can culture ask of AI?

— It's so new that I find it difficult to think of what to ask of you. It's like when electricity replaced steam. First there was technology, and then the reorganization of work and the production model. The question we must ask ourselves is what reconfiguration these tools allow us to do. There's a very interesting project, for example: augmented committees. Around a conflict, interviews are conducted with different interest groups and virtual opinions are formed. You have four or more perspectives at once.

And does AI also come to a conclusion?

— No, the conclusion must be human. It's important because it offers perspectives you might have missed and facilitates dialogue between perspectives, but the conclusion must be human.

Can't AI replace everything?

— That's not the goal. It's a mistake to delegate our intelligence and responsibility to AI. We must be careful not to fall into intellectual sedentarism. It's important to be the driver and for AI to be the co-pilot.

Should AI be copyrighted if it generates original content?

— This isn't resolved in the EU. It's important to consider this from a public perspective, both regarding copyright and originality. What worries me is the lack of responsibility shown by these companies and the lack of demands placed on AI users. It's an anecdote, but the other day Elon Musk's artificial intelligence started generating responses about a supposed "white genocide" in South Africa. No matter what question you asked, the answer was always that one. This demonstrates the danger of relying on mechanisms over which we have no control. With digital technology, we have socially refrained from making certain demands.