The EU AI Act’s Enforcement Machinery Is Now Operational—Here’s What Changed

The EU AI Act just crossed a critical threshold. In early June 2026, the European Commission appointed the Scientific Panel and Advisory Forum—the governance bodies that will oversee AI regulation across the EU. This isn’t procedural bureaucracy; it’s the moment when enforcement infrastructure shifted from assembly mode to operational readiness.

For anyone building or deploying AI in Europe, this timing matters urgently. The broader AI Act requirements take full effect on August 2, 2026—just two months away.

What’s Happening Right Now

The Commission’s transparency consultation closed on June 3, 2026, marking the final comment window on Article 50 rules for synthetic content labeling and watermarking. This feeds directly into the August deadline for transparency obligations, which require providers to mark AI-generated images, audio, video, and deepfakes in human-readable and machine-detectable formats.

In parallel, the Digital Omnibus—the first amendment package to the AI Act since June 2024—is moving toward formal adoption. This update introduces meaningful changes: staggered compliance deadline deferrals, new prohibitions on non-consensual intimate deepfakes and child sexual abuse material (effective December 2, 2026), and clarified interaction between the AI Act and product safety regulations like the Machinery Regulation.

Why This Matters for Builders

The operational activation of the AI Office’s governance infrastructure signals that enforcement is shifting from guidance to supervision. The Scientific Panel and Advisory Forum now have authority to advise on high-risk system classification, interpret transparency obligations, and support Member States in implementing national competent authorities.

For European AI developers and deployers, this means three things:

1. Transparency is imminent. Systems placing synthetic content on the market must comply with watermarking requirements by August 2, 2026. If your system was already on the market before that date, you have until December 2, 2026 to implement machine-detectable marking—but that grace period is tighter than originally planned.

2. High-risk classifications are being clarified. The Commission published draft guidelines on May 19, 2026 for classifying high-risk AI systems. These guide interpretation of intended purpose through instructions for use, marketing materials, and documentation—not just terms of service. Providers cannot escape high-risk classification by excluding uses in legal terms if product positioning tells a different story.

3. Standards and support tools are still in flux. The Code of Practice on AI-generated content marking was expected to be finalized in late June 2026. Harmonized standards for high-risk AI system requirements are likely in H2 2026 or H1 2027. This timing squeeze—enforcement approaching while guidance is still being published—creates real pressure on compliance teams.

Practical Implications

If you operate an AI system in the EU or target EU users:

  • Audit your AI inventory now against the latest draft guidelines on high-risk classification
  • Map all AI features by risk level (banned, high-risk, limited-risk, minimal-risk)
  • For transparency obligations: implement watermarking and disclosure mechanisms before August 2, or prepare for a December 2 extension if your system predates the regulation
  • Document human oversight, logging, and post-market monitoring requirements
  • Review vendor dependencies and contractual obligations around AI compliance

Open Questions

Several elements remain unresolved:

  • When will the Digital Omnibus achieve formal adoption and publication in the Official Journal? The Council indicated “coming weeks,” but this directly affects implementation timelines.
  • Will harmonized standards publish before or after the August deadline? Late publication could force interim compliance strategies.
  • How will national competent authorities interpret high-risk classification in practice, particularly the “material influence on decision-making” exemption?

For European founders and enterprises, the message is clear: enforcement infrastructure is now live, deadlines are weeks away, and guidance documents are the rules that will be applied. The time for theoretical preparation has passed.

Source: European Commission