Council gives final approval to simplified AI Act rules

Background

The European Union governs artificial intelligence through a single broad rulebook known as the AI Act, which sets out how AI systems may be placed on the market, used and supervised. A text of that size needs periodic upkeep, and the EU has been running a simplification agenda aimed at making existing legislation lighter and easier to apply. The changes to the AI rules travelled through that agenda as part of an "Omnibus VII" legislative package ^[1]. The European Commission frames the same exercise differently: it describes the AI omnibus as part of its Digital Simplification Package, and records that the amendments were introduced to keep the rules clear, simple and innovation-friendly ^[2].

What happened

On 29 June 2026 the Council of the European Union gave its final green light to a new regulation that streamlines and simplifies the rules on artificial intelligence ^[1]. The Council is the institution in which member state governments sit, and its final approval is the stage at which a legislative file clears political decision-making. What emerged is an amending instrument: the Commission records the adoption of the AI omnibus in June 2026 and states that the amendments entered into force on 27 July 2026 ^[2]. The gap between the two dates is short, so the changes became operative soon after the political agreement was sealed.

Key details

Two features stand out in the published record. The first is the direction of the exercise. Both institutions describe it as streamlining and simplifying ^[1][2] — trimming redundancy and ambiguity in the existing text rather than layering on new procedural obligations. The stated aim is a rulebook that stays clear, simple and innovation-friendly ^[2].

The second feature is a substantive prohibition written into the new law. According to the Council, a key provision prohibits AI practices involving the generation of non-consensual sexual and intimate content, or of child sexual abuse material ^[1]. The ban is framed around what the system does: AI used to produce sexual or intimate imagery of a person without that person's consent, and material recording the sexual abuse of children, falls within the prohibition. Naming these uses in the AI rules ties a category of content-related harm to the technology that produces it, and not only to the people who pass the material on.

Why it matters

The file is a useful illustration of what simplification means in practice. It is not deregulation in the sense of removing obligations wholesale; it pairs a lighter, tidier rulebook with at least one explicit ban. For companies building or deploying AI systems in the EU, a shorter and clearer text lowers the cost of working out which rules apply to them, which is the reasoning behind the stated goal of keeping the rules innovation-friendly ^[2]. For regulators, consolidation makes the rulebook easier to apply consistently across member states.

There are limits to what the record shows. The Council's account is an outline, not the operative text, so the precise scope of the prohibition and the way it is enforced will be settled by the regulation as adopted and by the authorities applying it. What the record does establish is a pattern that is likely to recur: the EU is treating simplification and safeguards as things that can travel in the same legislative vehicle, rather than as a trade-off in which one must wait for the other.