Key Developments

AI governance has shifted decisively from planning to execution in 2026, with the EU AI Act’s enforcement deadlines creating immediate compliance pressures across European organizations. The prohibited practices provisions, which took effect in February 2025, are now seeing their first enforcement actions as national regulators build inspection capabilities.

Critical upcoming milestones include the 2 August 2026 deadline for high-risk AI systems in the financial sector, with remaining provisions becoming fully applicable by August 2027. However, the current draft of the Digital Omnibus may extend these timelines, with long-stop dates potentially pushed to December 2027 for high-risk systems and August 2028 for product-embedded systems.

The European Commission has also published the first draft of the Code of Practice on Transparency of AI-Generated Content, introducing an “EU common icon” to help users identify AI-generated deepfakes and synthetic media at a glance.

Industry Context

What began as a legal compliance exercise has evolved into fundamental organizational transformation. In 2024, AI governance was typically a side responsibility for legal teams. By 2026, it has become a dedicated function, with many organizations appointing Chief AI Officers (CAIOs) to manage AI strategy and governance—a role now mandated for some public sector organizations under the EU AI Act.

The shift reflects the reality that governing agentic AI systems requires entirely new approaches compared to traditional software governance. Enterprise procurement processes have adapted accordingly, with security questionnaires now including AI-specific sections and RFPs requiring model cards and evaluation reports that didn’t exist months ago.

Practical Implications

For Irish and European organizations deploying AI systems, the compliance reality is immediate. Companies must now provide documentation about system behavior, evidence of testing protocols, and clear governance frameworks. The focus has shifted from theoretical policy compliance to operational tools and real-world implementation.

Organizations shipping AI applications are finding that customer expectations have fundamentally changed, with procurement teams expecting comprehensive documentation and governance evidence as standard practice.

Open Questions

Key uncertainties remain around the EU Cloud and AI Development Act, expected in Q1 2026, and how the Digital Omnibus will ultimately affect compliance timelines. The effectiveness of the new transparency measures for AI-generated content also remains to be tested in practice.

Source: artificialintelligenceact.eu