The European Union has agreed to simplify its landmark Artificial Intelligence Act, just under two years after it became law. EU governments and Parliament lawmakers struck a deal on May 7 that extends compliance deadlines, narrows the scope of high-risk AI systems, and reduces regulatory overlap.
The changes come through what’s called the “AI omnibus” – a package of targeted amendments designed to cut red tape while keeping the law’s core risk-based framework intact. Whether this represents smart course-correction or quiet deregulation depends on who you ask.
Key changes and extended timelines
The most significant change involves timing. High-risk AI systems covering employment, education, and health insurance now have until December 2, 2027 to comply – pushed back from summer 2026. AI embedded in physical products like medical devices gets even more time, with obligations delayed until August 2028.
The definition of “high-risk” has also been narrowed. Only AI systems whose failure would create genuine health or safety risks face the heaviest obligations. Tools that assist users or optimize performance no longer automatically trigger the full regulatory regime.
Other key changes include:
- Elimination of overlapping assessments where sector-specific laws already regulate AI functions
- Complete removal of machinery from the AI Act, now governed by separate sector-specific regulation
- New prohibition on AI tools that generate non-consensual sexually explicit images, including deepfakes
- Simplified technical documentation for small and mid-sized companies
Industry pushback and political concerns
The machinery carve-out came after heavy lobbying from companies like Siemens and ASML. But Green MEP Sergey Lagodinsky, a key voice on digital regulation, sees this as problematic.
“By having excluded machinery, we’re making a first step into fragmenting AI regulation,” he said. He pointed to the United States as a cautionary example, where the absence of clear federal guidelines has created a confusing patchwork of conflicting state rules.
“On the one side, everyone is saying how free and unregulated the US market of AI is. And on the other side, many don’t know which state regulates how,” Lagodinsky explained. “There are no clear guidelines and guardrails.”
What this means for companies
For businesses, the changes offer breathing room. Small and medium enterprises benefit from simplified documentation requirements, extended deadlines, and broader access to regulatory sandboxes where they can test AI systems under relaxed rules.
The approach remains proportional – a small company using an off-the-shelf chatbot faces far less regulatory burden than one selling high-risk AI for hiring decisions. But compliance still brings real costs, and fines remain for those who fall short.
Lagodinsky said he can accept the overall outcome. “The final agreement is something that we can be okay with. I do not belong to those who say this is a catastrophe.” However, he warned against constantly reopening legislation. “We cannot constantly reopen the legislative process and try to take shortcuts.”
The bigger challenge ahead
The changes highlight a fundamental question: can any law keep pace with rapidly advancing AI technology? Lagodinsky was direct about this challenge. “I am concerned that our legislative processes are much slower than the fast pace of innovation.”
He called on the EU AI Office and European Commission to act as “regulators-in-between,” filling gaps through guidance, codes of conduct, and enforcement action faster than full legislative cycles allow. “The commission is sometimes very timid or slow or late on acting, and that’s why it’s even more important that the commission and AI Office take their responsibilities even more seriously.”
The AI Act remains the world’s most comprehensive AI law even after these changes. Its risk-based framework stays intact. But the deal sets a precedent that the rulebook can be reopened. The next test is whether enforcement agencies will robustly implement what remains, or whether delays and carve-outs will quietly weaken the law’s original ambitions.
Formal approval by EU governments and the European Parliament is expected in the coming months.
