════════════════════════════════════════════════════════════════ AIDRAN STORY ════════════════════════════════════════════════════════════════ Title: Friedrich Merz Wants Industrial AI Exempted From EU Rules. Scholars Are Already Pushing Back. Beat: AI Regulation Published: 2026-04-20T22:16:45.193Z URL: https://aidran.ai/stories/friedrich-merz-wants-industrial-ai-exempted-eu-29a5 ──────────────────────────────────────────────────────────────── At Hannover Messe, German Chancellor Friedrich Merz made the case that industrial AI should be carved out from proposed EU rules, arguing that regulatory friction was costing Germany its competitive edge.[¹] The argument itself isn't new — versions of it have been circulating in Brussels corridors for two years. What's different now is who's making it and how plainly. A sitting head of government, at one of {{entity:europe|Europe}}'s most prominent industrial showcases, is essentially asking Europe's most ambitious tech governance framework to make an exception for the sector it was arguably most designed to cover. The response from the academic community arrived quickly. Michael Veale and a cohort of scholars engaged directly with the Bluesky posts surfacing the story, and their objection wasn't simply that Merz was wrong — it was that the "innovation vs. regulation" frame was a false one to begin with.[²] That framing dispute is where {{beat:ai-regulation|the regulation conversation}} is most alive right now. An op-ed circulating from Newsweek argues that asking whether governance hurts innovation is the wrong question entirely — the more productive question is what the innovation is for, and for whom. It's a reframe that keeps appearing in policy-adjacent spaces, but it hasn't yet broken through to the level where chancellors give speeches. Meanwhile, the EU AI Act's compliance calendar is pressing harder on the private sector than on governments. Law firms without AI governance policies face real exposure as Colorado and the EU's own rules both take effect in 2026. A separate thread of discussion — quieter but persistent — questions whether the compliance tooling that's proliferating is actually fit for purpose. One post drew pointed ridicule for AI governance infographics that dress up standard ISO 27001 security controls as novel LLM guidance, treating identity and access management as some kind of revelation. The frustration behind that mockery is real: the gap between what {{story:europes-ai-rulebook-real-enforcing-another-d2a7|the EU AI Act requires}} and what existing audit frameworks can actually assess is significant, and the industry is paper-filling that gap rather than closing it. There's a parallel conversation running on the governance-versus-exemption axis at the global level, too. The framing at {{entity:india|India}}'s AI Summit, per observers, has shifted away from containment arguments and toward managing the reality of AI's proliferation — an acknowledgment, however tacit, that the window for precautionary governance may have already passed. The US posture, as several posts note, is increasingly about preventing other countries from setting the rules: keeping allies aligned against frameworks that might constrain American AI development. {{story:trumps-ai-policy-contradiction-built-discourse-14e5|That deregulatory push has its own contradictions}}, and they're not getting quieter. What the Merz exemption push actually signals is a stress fracture in the EU's approach that has been visible for some time: member states built the AI Act together and are now individually lobbying to be left out of it. If industrial AI gets a carve-out, the question immediately becomes which sectors don't qualify — and the answer to that question will be negotiated by people who also have productivity targets to meet. The scholars pushing back on Merz aren't wrong. They're just arguing with someone who isn't listening to the same incentives they are. ──────────────────────────────────────────────────────────────── Source: AIDRAN — https://aidran.ai This content is available under https://aidran.ai/terms ════════════════════════════════════════════════════════════════