Ramón López de Mántaras, a pioneer of artificial intelligence in Spain and founder of the Artificial Intelligence Research Institute of the CSIC (IIIA-CSIC), has been on the front line of AI for decades. But far from celebrating the current boom in generative AI, he warns of its serious consequences: unsustainable energy consumption, loss of cognitive abilities, and corporate dominance that threatens digital sovereignty. In an interview on the occasion of the Honorary Collegiate Award from the Professional College of Computer Engineering of Galicia (CPEIG), López de Mántaras defends an ethical, humanistic AI at the service of the common good.

The beginnings of AI: a dream not yet fulfilled

López de Mántaras began working in AI in 1975, when the field was an academic niche without corporate presence. "There were few of us," he recalls. "Already in the 1960s, it was said that in 15 or 20 years we would have artificial intelligence comparable to human intelligence. We are in 2026 and that is not reality." The goal of emulating human intelligence remains the same, but it is now in the hands of large technology corporations, which has radically changed the course of research.

Generative AI: too high a price

For López de Mántaras, generative AI represents a dangerous detour. "The AI that was done more than four years ago was more cognitively interesting," he states. "Now we pay a huge price: critical, unsustainable energy consumption. We are talking about 5 or 10% of the energy consumption of a large country, plus water consumption." The researcher criticizes that classical AI, which achieved extraordinary results at a much lower cost, has been abandoned. "There are alternatives, such as neuro-symbolic AI, which combine the best of both worlds, but the big technology providers have the upper hand and do not want to change."

The path to artificial general intelligence (AGI) through generative AI is, in his opinion, a mistake. "We should reorient generative AI towards integration with classical AI, solving specific problems instead of seeking an AGI that, moreover, generates hallucinations and excessive consumption."

Ethics and red lines: not everything goes for business

López de Mántaras is blunt: "Applications that cede to machines aspects that impact people, such as lethal autonomous weapons, mass surveillance of citizens, or information manipulation, should not be made. All red lines have already been crossed for business." Even companies like Anthropic, which present themselves as ethical, practice "ethic washing." To stop this, he proposes two avenues: regulation and education. "We must train engineers and scientists in ethics, and give more weight to the Humanities in the education of young people. We are digitizing poorly, public administration is an example."

Regulate and break up big tech

The researcher calls for strong measures: "Countries must have the courage to confront big tech, which has become too big. They must be forced to break up so they lose power." He criticizes that the European Union has yielded to tech lobbies, delaying the AI law until 2027 and avoiding the obligation to publish water and energy consumption. "Europe is rolling out the red carpet for them."

In this context, digital sovereignty becomes crucial. López de Mántaras points out that dependence on tools like ChatGPT can lead to a loss of cognitive abilities. "Writing is thinking. When you delegate to ChatGPT, you delegate your ability to think. Studies already show the impact on students." He also warns about the risk of being trapped: "If OpenAI raises prices or the network goes down, companies that have delegated all their development to these tools will have to pay millions in licenses or will not be able to operate."

The Luddite legacy: a call for reflection

Far from being a technophobic stance, López de Mántaras reclaims Luddism as a warning: "Progress is not neutral. The Luddites were not opposed to the future, but to a future built without counting them. It is a call to have a critical view on the use of technology." At a time when generative AI advances unchecked, his message resonates strongly: we need an AI that contributes to knowledge, social progress, and collective well-being, not to the profit of a few corporations.

For IT companies and professionals, López de Mántaras's warnings are a call to action. Data resilience and technological sovereignty are key issues in the age of AI, as highlighted in articles like Veeam strengthens its channel ecosystem with Michelle Graff or Success Story: Ethical Hacking. Digital transformation must be ethical and sustainable, as explored in this logistics success story. Even transparency in social networks, such as Elon Musk's opening of X's code (read more), is a step in the right direction, albeit insufficient.

Original source: ComputerWorld. Analysis and adaptation by ForgeNEX.