European Commission mulls AI Act delays in face of Trump and business pressure

"The European Commission is considering plans to delay parts of the EU's landmark Artificial Intelligence Act, after intense pressure from businesses and Donald Trump's administration. The commission confirmed that a reflection is still ongoing on delaying aspects of the act, after media reports that it was weighing changes to the law with the aim of easing demands on companies. The EU's act, the first comprehensive legislation in the world regulating artificial intelligence, came into force in 2024, but many of its provisions do not yet apply."

"According to the Financial Times, the commission is considering giving a one-year grace period to companies breaching the rules on the highest-risk AI. Providers of generative AI systems that can produce content, such as text or images who have already placed products on the market before the implementation date could be granted a one-year pause from the laws to provide sufficient time to adapt their practices within a reasonable time without disrupting the market, stated an internal document cited by the FT."

"The commission is also considering delays to imposing fines for violations of its new AI transparency rules until August 2027 to provide sufficient time for adaptation of providers and deployers of AI systems to implement the obligations, the paper reported. Also being studied is greater flexibility for AI developers of high-risk systems over monitoring performance of products on the market, by allowing them to follow guidance that would be less prescriptive than the system originally envisaged, according to MLex, which first reported on the planned amendments to the act."

The European Commission is reviewing proposals to delay parts of the EU Artificial Intelligence Act in response to pressure from businesses and the US administration. The Act entered into force in 2024, but major obligations for high-risk systems are not due until 2026 or 2027. Considerations include a one-year grace period for some providers that already placed generative AI products on the market, delaying fines for transparency rule breaches until August 2027, and allowing less prescriptive guidance on post-market performance monitoring for high-risk developers. Proposals are expected by 19 November and could change before publication.

Read at www.theguardian.com