- ■
The European Parliament voted to postpone compliance with EU AI law for high-risk systems until December 2027, and sector-specific regulations until August 2028, according to Parliament’s official record.
- ■
The same vote approved a complete ban on the nudify app, which targets AI tools that generate non-consensual intimate images.
- ■
Companies operating in the EU, such as Google, Meta, Microsoft, and OpenAI, now have more leeway to implement safety controls for AI systems that are considered to pose serious risks.
- ■
This dual move signals regulators’ intention to expedite banning harmful applications while extending technical compliance timelines.
European lawmakers have just rewritten the world’s most ambitious AI regulation timeline. The European Parliament overwhelmingly voted to extend the EU AI law compliance deadline until December 2027, giving companies building risky AI systems nearly two more years to meet the requirements. However, this delay has its pitfalls. Lawmakers also moved to ban nudity apps completely, showing no compromise on AI systems that weaponize deepfakes against individuals.
The European Parliament just gave AI companies a compliance reprieve while closing the door on one of the technology’s most fraudulent applications. In the vote, which passed with wide support, lawmakers extended a key deadline for EU AI legislation, but also moved to ban nudity apps that use AI to create intimate images without consent.
This shift in timing is significant for companies competing to comply with regulations that have become the global template for AI regulation. High-risk AI systems deemed to pose a “significant risk” to health, safety or fundamental rights must meet the law’s requirements by December 2027. This is a significant extension from the previous deadline, when Google, Microsoft, Meta and OpenAI were busy overhauling their systems relating to EU users.
But in certain areas it gets even more complicated. Companies developing AI systems that are subject to existing safety regulations (such as medical devices, toys, and automotive applications) will not need to comply until August 2028. The European Parliament’s official press release confirms that this phased approach reflects the technical complexity of refurbishing AI systems with the transparency and safety controls required by law.
However, the delay is not unlimited. Rules requiring AI providers to watermark synthetic content and other basic transparency measures are still moving forward on their original schedule. And the banning of the nudify app shows that regulators are not taking back what they believe is clear harm.
