The European Parliament has given final approval to amendments to parts of the EU AI Act, as part of a broader “digital omnibus” package. The vote is a sign that lawmakers want to adjust how some AI rules are timed and enforced, while also drawing a hard line on certain harmful uses.

In the roll-call vote, 423 members voted in favor, 57 against, and 174 abstained. That split suggests there was enough disagreement to keep the discussion politically active, even after final approval.

According to the Parliament’s announcement, the package includes postponing some obligations under the AI Act. For people who will be affected by compliance requirements, that kind of delay can matter, because it changes when businesses and other groups actually have to meet certain legal steps.

Ban targets nudifying and related abuse

The same package also includes an EU ban on “nudifying tools” and on AI-assisted creation of child sexual abuse material. The Parliament’s wording puts these items in the category of conduct the EU intends to prohibit, not just regulate.

Why this is important for everyday life is fairly straightforward: tools that generate intimate or sexual content without consent, and tools that help produce child sexual abuse material, are exactly the kind of misuse that can cause real harm quickly and at large scale. A ban aimed at these specific capabilities is meant to make enforcement clearer, even as other AI Act obligations are being shifted in time.

The Parliament’s decision lands as the EU continues to refine the AI Act through packages rather than in one single move, which can make the final compliance picture feel like it’s changing while organizations are trying to plan ahead.

Source: European Parliament