EU AI Act Enforcement Framework Now Operational

The European Commission has taken a major step toward operationalizing the EU AI Act by appointing a 60-member Scientific Panel and establishing a complementary Advisory Forum. These bodies, with members serving two-year terms starting June 1-4, 2026, mark the first concrete enforcement infrastructure since the AI Act’s adoption in June 2024.

Key Developments

The Scientific Panel brings together world-leading independent experts spanning frontier AI development, technical auditing, industry practice, and societal impact assessment. Their mandate focuses specifically on general-purpose AI (GPAI) models, systemic risks, model classification, evaluation methodologies, and cross-border market surveillance—essentially the technical heavy lifting needed to make the AI Act work in practice.

The Advisory Forum complements this with broader expertise from academia, civil society, and industry (notably including SMEs and startups), addressing standardisation challenges and implementation hurdles across the full scope of the regulation.

Simultaneously, negotiators from the Council, Parliament, and Commission reached a provisional agreement on the Digital Omnibus on AI—the first amendment package to the original Act. Most notably, two new prohibited AI practices now take effect December 2, 2026: generating non-consensual intimate material and child sexual abuse material (CSAM).

Why This Matters

Enforcement infrastructure has been the missing piece in AI governance. Having expert advisors embedded in the Commission creates the technical capacity to classify AI systems correctly, evaluate novel risks, and make evidence-based decisions. Without this, the AI Act risks becoming bureaucratic theatre.

For Ireland specifically, this represents a parallel development: the Irish government published the General Scheme of the Regulation of Artificial Intelligence Bill 2026 on February 4, establishing the legislative framework to implement the EU Act domestically. Most significantly, Ireland will establish its own AI Office by August 2026, serving as the central coordinating authority for supervision, enforcement, and support for safe AI deployment across the economy.

This is Ireland’s opportunity to position itself as an EU leader in AI governance, particularly given the country’s existing role as a tech hub.

Practical Implications

For AI builders and developers: The Scientific Panel’s work on model classification and evaluation methodologies will directly shape how your systems are assessed. Expect clearer (but still evolving) guidance throughout June-July on what “high-risk” means in practice.

For Irish organisations: The August 2026 AI Office establishment creates a domestic touchpoint for compliance queries and enforcement decisions. It’s worth monitoring the Bill’s progress through the Oireachtas for implementation details.

For users and civil society: The prohibited practices amendments on intimate imagery and CSAM represent concrete protections, with enforcement beginning this December.

Open Questions

  • How quickly will the Scientific Panel publish practical guidance on GPAI classification?
  • What will Ireland’s AI Office’s relationship be with the EU Commission’s enforcement teams?
  • How will the May 19 draft guidelines on high-risk classification evolve following the June 23 public consultation deadline?
  • Will SME compliance support keep pace with enforcement requirements?

The infrastructure is finally being built. Now comes the execution phase.

Source: European Commission & Irish Department of Enterprise, Trade and Employment