The European Commission is considering plans to postpone parts of the EU’s artificial intelligence law following intense pressure from businesses and President Donald Trump’s administration.
Following media reports that the city of Brussels was considering changes to ease demands on businesses, the commission confirmed that “reviews” of delayed aspects of the regulations were “still ongoing”.
The world’s first comprehensive law regulating AI took effect in August 2024, but many of its provisions have not yet been applied. Most of the obligations on companies developing high-risk AI systems that “pose a significant risk to health, safety or fundamental rights” are not expected to come into effect until August 2026 or a year later.
According to the Financial Times, the commission is considering giving a one-year “grace period” to companies that violate the most risky AI rules. An internal document cited by the FT states that providers of generative AI (systems that can generate content such as text or images) that have already brought products to market before the effective date may be given a one-year reprieve from the law “to give them sufficient time to adapt their practices within a reasonable period of time without disrupting the market.”
The commission is also considering postponing the imposition of fines for violations of the new AI transparency rules until August 2027 to “ensure sufficient time for providers and adopters of AI systems to adapt” to meet the obligations, the newspaper said.
It is also being considered to give AI developers of high-risk systems more flexibility when it comes to monitoring the performance of products on the market, by allowing AI developers to follow less prescriptive guidance than the systems originally envisioned, according to specialist news site MLex, which first reported on the proposed changes to the law.
This proposal is subject to change before the planned release on November 19th. Once published, it will require the consent of EU member states and the European Parliament.
The EU has come under repeated pressure from the Trump administration to weaken regulations on technology companies. The U.S. president recently threatened to impose tariffs on countries that have technology regulations or digital taxes “designed to harm or discriminate against U.S. technology.”
Meta announced earlier this year that it would not sign the Commission’s Code of Practice on General-Purpose AI Models. “Europe is on the wrong path when it comes to AI,” wrote Joel Kaplan, the company’s chief international affairs officer, arguing that the code creates “legal uncertainty” for model developers and introduces measures “far beyond the scope of AI law.”
But American companies are not alone in their frustration with European regulations on rapidly evolving technology.
After newsletter promotion
Dozens of European companies are calling for the law to be suspended for two years to allow time for “reasonable implementation” and “further simplification of the new rules.” The open letter, signed by the heads of 46 companies including Airbus, Lufthansa and Mercedes-Benz, said such delays would show innovators and investors that Europe is “seriously committed to the challenge of simplification and competitiveness.”
“The possibility of delaying the implementation of the targeted parts of the AI Act is still under consideration,” European Commission spokesman Thomas Renier said in a statement. He said no decision had been made, adding that the committee “will always fully support the AI Act and its objectives.”
Mr Rainier said Brussels was “in constant contact with our partners around the world”, but said it was not for third countries to decide how the EU would legislate. “This is our sovereign right,” he added.
