Very discreetly, on the 19th of this month, the European Commission issued a press release announcing a major change to the regulatory package for artificial intelligence. Press release, that’s it. Hours later, a Commission official came to speak to reporters and try to dispel any doubts. There was no one holding a high position in the bureaucracy. Whoever wrote about the impromptu press conference did not even mention his name. In this way, Europe informed the world that strict rules for “high-risk” uses of artificial intelligence would not begin in August 2026. But they were postponed until December 2027. The European Union backed down.
- And set for 2026: What is the motivation behind the division between Bolsonaro and Michel’s children?
When the AI law was approved in March 2024, Commission President Ursula von der Leyen gave a speech. She was followed by the Vice-President, then the European Internal Market Commissioner. They wanted the headlines and they got it.
- The state is in crisis: Correos expects the first installment of the loan this year
– The first comprehensive regulatory framework for AI systems worldwide – said von der Leyen.
The model will establish a global standard, as many analysts bet. (Including this one.)
What happened is not difficult to explain. Two years later, Donald Trump became President of the United States and put his foot on the AI accelerator. The Americans treat China as their main adversary, and are in a race to see who can first create artificial intelligence capable of matching human intelligence. It’s not certain that such technology is possible, but no one wants to bet that it won’t happen. Rules may be good to protect citizens. But with the US and China being so aggressive, offering all manner of incentives to the sector, the Europeans have made the obvious calculations. If there are no rules in one corner and the bureaucracy is extensive in Europe, those who know how to build artificial intelligence go to China or the United States. The continent has been excluded from the game. Then it blinks.
The European base is good. I have identified some levels of distinct uses of AI. In the unacceptable risk category people are manipulated using technology; Classifying residents by biometric characteristics such as race, sexual orientation, or religious belief; Using artificial intelligence for real-time monitoring. This was banned as of February this year, and remains banned.
In the high-risk category, artificial intelligence is used to draw conclusions about someone’s health; using artificial intelligence to identify areas where police need to take greater action; Use of public data to identify individuals – electoral registers, documents in registration offices. Even using personal data from European citizens for AI training was in this high-risk category, which would have many requirements to allow it. These rules will come into effect at the beginning of the second half of next year. Because everyone got an extra year and a half to work in their spare time.
Not surprisingly, no European authority wants to make a big announcement about this change. The blow came from several civil rights organizations immediately. Also from newspapers. The criticism was clear in tone: the European Union had bowed to the power of Big Tech.
But the consequences are not limited to Europe. If the European Union cannot regulate, neither Latin American nor Asian countries will be able to. In a way, the Old Continent has provided shelter to those who claim that stricter rules are necessary. After all, the common market region’s domestic GDP is higher than that of China. Big Tech could not have chosen to leave the market. They might not produce there, and they might have tools suitable for Europeans, but getting away was very difficult. This security served as a guarantee for any other country that decided to follow the same path.
no more. In essence, the Trump presidency changed the course on which the world was heading. With fewer rules, the sector is certainly freer to develop. This represents acceleration, and more artificial intelligence will emerge, capable of doing more things. It is not absurd to think that they will bring countless benefits. But it is certainly difficult to predict all the consequences. It is not an inexpensive option. After all, the world’s population has no vote in the decisions that will be made using the information we all collectively produce every day.