In late April to early May 2026, a major shift occurred in EU AI governance: EU legislators agreed to delay high-risk AI rules while introducing an AI Act "Omnibus Simplification Package."
The news has triggered two conflicting interpretations: one camp sees "regulatory relief," while the other warns that "rules are more complex, just in a different form."
What Happened
- August 2024: EU AI Act enters into force, with phased implementation
- February 2025: High-risk AI system rules begin application
- April 2026: Commission proposes Omnibus simplification — delaying high-risk rules, reducing SME burden, narrowing AI definition
- Early May 2026: Member states and lawmakers fail to reach final deal on details (Reuters), but direction is set
The Challenge: A Regulatory Rebalance, Not a Retreat
The IAPP analysis identifies three adjustments: what is AI, who needs oversight, and how to enforce.
- Narrower AI definition — excluding rule-based systems, focusing on systems with genuine autonomous decision-making
- Slower implementation — extended deadlines for high-risk AI compliance
- Clearer liability chain — better defined responsibilities between foundation model providers and deployers
What This Means for Enterprises
Good news: More transition time (especially for finance, healthcare, HR AI applications); ~40% reduced documentation requirements; broader SME exemptions.
Challenges: "Simplification" ≠ "simple" — companies need to re-understand AI classification standards; compliance paths have changed but the destination hasn't; enforcement divergence across member states.
Convergence with China's Approach
Interestingly, the EU's simplification direction converges with China's regulatory thinking — both moving from "blanket regulation" toward "risk-classified management." For cross-Eurasia enterprises: core compliance elements are converging (transparency, risk management, human oversight), but specific paths and language differ.
Looking Ahead
JD Supra's assessment is apt: "The Omnibus simplification isn't the end of the AI Act — it's its coming of age." AI governance is shifting from a "legislative paradigm" to an "enforcement paradigm." The real test is how rules are interpreted and enforced in real business environments.
For Chinese and multinational companies deploying AI: don't wait for final clarity — start building AI governance frameworks now; risk classification is central; human-in-the-loop is evolving from recommendation to requirement.
References
- "EU AI Act Omnibus Simplification Proposal" — European Commission, April 2026
- "EU lawmakers fail to reach final deal on AI Act simplification" — Reuters, May 2026
- "EU AI Act Omnibus: The IAPP Analysis" — International Association of Privacy Professionals (IAPP), April 2026
- "EU AI Act Omnibus Simplification: A Coming of Age" — JD Supra, May 2026
💡 Did this article spark an idea for your work?
humanaifit explores how humans and AI truly collaborate. If you face real challenges in enterprise AI adoption, human-AI fit, or global compliance, we'd love to hear from you.
🔗 Join the AI Era Survival Guide Knowledge Planet — where every article comes with practical toolkits, templates, and direct discussions with the authors.