The enforcement machinery has human faces now.
On June 1, the European Commission formally appointed members to the EU AI Act’s Scientific Panel and its Advisory Forum, the two independent expert bodies that support the European AI Office in monitoring and enforcing compliance with the Act’s most demanding obligations. The appointment is procedural in form. It’s structural in consequence.
According to the Commission’s announcement, named scientific experts include Adam Gleave (UK), Cassidy Nelson (Australia), and Anna Katariina Wisakanto (Finland). ⚠️ These names are drawn from a single official source and require human validation before publication. Permanent seats on the Advisory Forum are held by the Fundamental Rights Agency (FRA), the European Union Agency for Cybersecurity (ENISA), and the European standards bodies CEN, CENELEC, and ETSI, institutional membership consistent with the governance structure established in Articles 67 and 68 of the EU AI Act itself.
The Scientific Panel’s primary role is technical: it provides independent expert review of GPAI model compliance, evaluating whether foundation models meet the Act’s systemic risk thresholds. The Advisory Forum sits one layer above, drawing in civil society, industry, and standards bodies to advise the AI Office on enforcement priorities and guidance development. Neither body issues fines directly. Both feed the enforcement judgments of the AI Office, which does.
Who This Affects
The June 23 deadline is the immediate action item
The AI Office’s public consultation on high-risk AI classification guidelines under Article 6 closes June 23. That’s the window for compliance teams and developers to submit feedback that shapes how “high-risk” is defined before the classification criteria are locked. The bodies appointed June 1 will be the ones reviewing that feedback. Organizations that want to influence the framework now have a named audience and a 22-day deadline.
The catch is that the consultation window doesn’t pause compliance work. Annex III high-risk system developers are building compliance programs toward a December 2027 deadline, a deadline that’s now being set against guidance that’s still forming. Submitting feedback before June 23 isn’t just a policy participation exercise. It’s a direct input into the standards that will govern your own audit requirements.
The real question is what comes out of the consultation. The EU AI Act’s Article 6 classification criteria determine which AI systems face the Act’s most burdensome obligations: conformity assessments, technical documentation, human oversight requirements, and registration in the EU database of high-risk AI systems. If the guidelines narrow the definition, some systems currently scoped for high-risk compliance may fall out. If they expand it, compliance programs that looked manageable may face a harder scope. The Scientific Panel and Advisory Forum are now the bodies that will weigh in on that call.
Analysis
The June 23 consultation window is not a formality. The Scientific Panel appointed June 1 will review submissions from this consultation. Organizations that submit specific, technically grounded feedback on Article 6 classification criteria are directly influencing the standards that will govern their own compliance requirements.
Enforcement in the EU has been building piece by piece since February 2025. The prohibited practices ban took effect then. GPAI model obligations followed in August 2025. The compliance machinery activation in May set the broader infrastructure in motion. June 1 is the step that names the people doing the work.
Don’t expect the Scientific Panel to make noise immediately. It takes time for newly appointed expert bodies to develop working methods, internal procedures, and enforcement positions. But the appointments remove the last structural excuse for treating EU AI Act GPAI enforcement as abstract. The Office has its experts. The consultation closes in 22 days. The timeline is now operational.