By continuing to browse our site you agree to our use of cookies, revised Privacy Policy and Terms of Use. You can change your cookie settings through your browser.
CHOOSE YOUR LANGUAGE
CHOOSE YOUR LANGUAGE
互联网新闻信息许可证10120180008
Disinformation report hotline: 010-85061466
A major provision of the European Union's AI Act is set to take effect on Sunday. The legislation will ban the use of artificial intelligence systems deemed to carry an unacceptable level of risk.
Under the EU's AI Act, some applications – like certain biometric categorization tools – will be banned. Failure to comply could result in fines of up to 7 percent of annual turnover.
Systems deemed to be high-risk include critical infrastructure such as transportation that could put lives in danger, the use of AI to filter CVs in job applications and its use in law enforcement to process evidence.
"These AI systems, they can still be put on the market but the makers need to make sure that their systems comply with some requirements that are in the Act and they need to either self-certify or get a third party to certify that they're in compliance with the act," said Pierre Larouche, Professor of Law and Innovation, University of Montréal.
Protecting rights while supporting innovation
The AI Act was adopted back in 2024, but the majority of its rules won't kick in until August 2026. The EU says it aims to protect people's rights while supporting an environment for innovation. According to Commission President Ursula von der Leyen, just 13.5 percent of EU businesses are currently using AI technology. She says that needs to change if the bloc is to remain competitive on the world stage.
"This year we will therefore launch a broad strategy for our continent. It will include and apply AI initiatives to drive industrial adaptation of artificial intelligence in key sectors," Von der Leyen said.
The recent attention around Chinese AI firm DeepSeek's new model has increased scrutiny on the sector. EU rules to ensure greater safety and transparency for general purpose AI tools are set to come into effect in a few months time.
"The AI office is indeed the responsible body for the supervision of these market developments when it comes to advanced AI models – so, including Deepseek, potentially. But these rules will only kick in in August 2025, later this year," said Commission spokesperson Thomas Regnier.
Advances in AI are moving fast. The challenge for the EU is establishing guardrails while not pulling the plug on a technology with so much potential.