EU Agrees to Delay High-Risk AI Restrictions Amid Industry Pressure

_{Photo credit: AntonKhrupinArt / Shutterstock.com}

The European Union has agreed to postpone its plans to introduce restrictions on high-risk AI systems in a significant delay to part of its AI Act.

The EU AI Act became law in August 2024, with restrictions on the use of high-risk AI originally planned to be introduced in August 2026. 

Under the Act, high-risk AI systems are those assessed as posing a potential risk to people's safety or fundamental rights. 

Following an agreement made by EU legislators early on Thursday, the planned restrictions for high-risk AI systems will now be delayed until December 2027. The agreement follows pressure from industry groups and several European countries, concerned that the bloc’s strict rules could weaken its global competitiveness in the AI race.

The US federal government, for instance, has not introduced comparable restrictions on the uses of AI to encourage innovation.

European Commission President Ursula von der Leyen said:

I welcome the political agreement on our Digital Omnibus on AI. This provides a simple, innovation-friendly environment for our European AI ecosystem to grow. At the same time, we are strengthening protections for our citizens. For safe and simple AI governance in Europe.

POLITICO reports that the agreement also “largely exempts” AI used in industrial applications from the AI Act’s scope. This change was strongly backed by Germany, which did not want major German engineering companies like Siemens to face overlapping requirements under both AI and machinery rules. 

On top of this, the agreement gives companies a three-month grace period to comply with new rules to add watermarks to AI-generated content. It also bans AI systems that generate sexualized deepfakes of identifiable people, as well as systems capable of producing child sexual abuse material.

In January, a cross-party group of 57 MEPs urged the European Commission to confirm that AI applications used to undress people without their consent were banned in the EU under the AI Act for violating people’s fundamental rights.