“La autorregulación no funciona.”
Self-regulation does not work, Pedro Sánchez told employers and union officials at the Moncloa Palace in Madrid on Monday, closing a government event on artificial intelligence hours before flying to New York for the UN General Assembly.
The prime minister said that recent statements from some key figures make it hard to know what is true, and suggested that part of their goal may be to attract investors, without naming anyone.
His comments came nine days after Dario Amodei published an essay urging the industry to slow down, which Sam Altman, Elon Musk, and Demis Hassabis supported on the same day.
He said he had read an interview where one of these “techno-oligarchs” asked for trust, but his answer was no.
He argued that having powerful, unregulated AI controlled by only a few people is a recipe for disaster. Also, he added that the current situation is not a race, but a stampede, using a phrase he heard from others in the room.
He ended by pointing out that some business leaders still claim regulation hurts innovation, just as they did 150 years ago.
This speech was given at the launch of Plan IA360, a twelve-month program that includes an AI gigafactory, new models developed with the Barcelona Supercomputing Center, a voucher scheme to help small businesses and the self-employed use AI.
A goal for more than half of Spanish companies to use AI by 2030, curriculum changes, and stricter oversight of advanced models.
When talking about data centers, he said that communities will judge them by their water and power use.
Next month, he will bring together social partners to begin talks on what he calls a new Social Contract for AI, focused on job security.
He also wants companies to be legally and financially responsible for what their algorithms do, including deepfakes, harm to young people’s mental health, and theft of intellectual property.
He said there is no need for a new international organization. He argued that the necessary structures already exist, such as the EU AI Office and the UN scientific panel on AI, which Spain helped negotiate and hosted in Madrid.
Now, he said, the focus should be on strengthening these efforts with partners like Canada, Japan, and other European countries.
Spain has achievements in this area, but also some gaps. It was the first country to establish a dedicated AI supervisory agency, AESIA, now based in A Coruña. However, its national AI law is still only a bill.
The cabinet sent it to parliament in May, but it is not expected to pass before the end of 2027. Until then, binding rules will come from EU regulations, not from Madrid.
He described the AI Act as a Spanish achievement, since it was agreed during Spain’s EU presidency.