EU member states and European Parliament lawmakers reached a provisional agreement in May 2026 to water down and delay key provisions of the bloc’s landmark AI Act, following nine hours of negotiations. The deal still requires formal endorsement by EU governments and the European Parliament in the coming months.
The most significant change pushes the enforcement deadline for high-risk AI systems — including those involving biometrics, critical infrastructure, and law enforcement — from August 2, 2026 to December 2, 2027. Machinery was also excluded from the Act’s scope, as it is already covered by separate sectoral rules.
Cyprus, which currently holds the rotating EU Council presidency, welcomed the outcome. Deputy Minister for European Affairs Marilena Raouna said the agreement “significantly supports our companies by reducing recurring administrative costs.”
The revisions are part of a broader European Commission push to streamline digital regulations after businesses complained that overlapping rules and administrative burdens were hampering their ability to compete with U.S. and Asian rivals.
The deal also introduces new consumer protections. A ban on AI practices that generate unauthorized sexually explicit images — a measure responding to such content produced by Elon Musk’s xAI chatbot Grok on X, including sexually intimate deepfakes — will take effect December 2. Dutch lawmaker Kim van Sparrentak said the ban means “everyone, but especially women and girls will be safe from horrific nudifier apps being widely available on the EU market.” Mandatory watermarking of AI-generated output will also apply from December 2.
The AI Act originally entered into force in August 2024, with provisions set to be enforced in stages. Despite the changes, the rules are still considered the strictest AI regulations in the world. Critics, however, argue the revisions reflect Europe yielding to pressure from major technology companies.
Source: Tech-Economic Times