Who Gets Protected, Who Gets Exposed
EU legislators on Thursday morning agreed to postpone restrictions on high-risk uses of artificial intelligence in the EU by more than a year, a move that hands industry and capitals a fresh victory while ordinary people are left to live with the consequences of rushed automation and weakened oversight. The deal also largely exempts the use of AI in industrial applications from the scope of the law, after negotiations that started on Wednesday evening and lasted until around 4.30 a.m. on Thursday.
The Cypriot presidency of the Council of the EU and the European Parliament confirmed the agreement after the overnight bargaining session, with the deal supported by European Parliament lawmakers and EU countries after heavy pressure from industry and capitals. That is the machinery at work: the people at the top negotiating the terms, the corporations applying pressure, and everyone else expected to accept the outcome as “governance.”
Industry Wins, Workers Wait
The change is a big win for Germany after top officials including Chancellor Friedrich Merz pushed for the change to keep tech heavyweights Siemens and Bosch competitive. EU countries had backed Germany’s demand to avoid a double regulatory burden for companies using industrial AI, and they will now only have to comply with AI requirements under separate machinery rules. Other industries under discussion, including medical devices, were not exempted and will still be covered by the AI law, negotiators confirmed.
The hierarchy is plain enough in the language of the deal itself: the burden is something companies must be spared, while the risks are something the public is expected to absorb. The rules were not removed for everyone, only bent where the pressure from powerful firms and governments was strongest.
What They Call “Innovation-Friendly”
The deal marks the first significant rollback of rules in the digital space, as the EU faces pressure from the U.S. over its tech laws and amid warnings from its own industry and governments that strict restrictions had put the bloc at a disadvantage in a global AI race. Commission President Ursula von der Leyen said it “provides a simple, innovation-friendly environment” for AI in Europe. “At the same time, we are strengthening protections for our citizens. For safe and simple AI governance in Europe,” she said on X.
That promise of “protections” comes alongside a delay of restrictions on high-risk AI until December 2027, more than a year later than planned. The phased rollout that was supposed to bring those rules in this August has been pushed back again, showing how quickly “safeguards” can be traded away when industry and governments decide the timetable is inconvenient.
The deal also gives companies a grace period on meeting new requirements to watermark AI-generated content, though that grace period will be three months rather than the six months originally proposed. Even here, the language is about easing compliance for companies first, with the public left to navigate whatever synthetic content floods the system in the meantime.
What Was Kept, and Why
The agreement does include a ban on AI systems that can generate sexualized deepfakes of intimate parts of “identifiable” people, after global outrage over the abusive use of Elon Musk’s AI tool Grok. AI systems that generate child pornography will also be banned.
The EU’s AI Act became law in August 2024 after years of talks. Under the phased rollout, rules governing high-risk uses were set to kick in this August. Instead, the latest deal pushes those restrictions to December 2027, while carving out a broad exemption for industrial AI and preserving separate rules for some sectors like medical devices.
What emerges is a familiar pattern of managed reform: a law announced as protection, then narrowed under pressure from the powerful, then repackaged as responsible governance. The people most exposed to the consequences of AI systems do not appear in the negotiation room, but they are the ones who will live with the delays, the exemptions, and the corporate-friendly fine print.