The European Union’s Artificial Intelligence Act reached its most consequential milestone today as high-risk AI system requirements, transparency obligations, and enforcement powers formally took effect across all 27 member states.
The provisions, which apply from August 2, 2026, require operators of high-risk AI systems to conduct conformity assessments, affix CE marking, and comply with detailed transparency rules under Article 50 of the regulation. The AI Office, the EU’s dedicated enforcement body, now has authority to oversee compliance and impose penalties for violations.
High-risk AI systems covered by the regulation include those used in critical infrastructure, education, employment, law enforcement, and migration management. Companies deploying these systems must implement risk management processes, maintain technical documentation, and ensure human oversight mechanisms are in place before placing products on the EU market.
The regulation has been phased in since February 2025, when prohibitions on certain AI practices and general AI literacy obligations first became enforceable. Governance rules and obligations for general-purpose AI models followed in August 2025, setting the stage for today’s broader enforcement push.
U.S. technology companies face particularly significant compliance burdens. Firms selling AI-powered products in Europe must now meet the same requirements as EU-based developers, regardless of where the technology was built. Holland & Knight legal analysts noted that most remaining provisions take effect today, with Article 6(1) obligations postponed until August 2027.
The enforcement timeline has not been without controversy. Industry groups lobbied for delays, arguing that conformity assessment infrastructure was not yet mature enough to handle the volume of AI systems requiring certification. EU officials countered that further postponement would undermine the bloc’s credibility as a global standard-setter for AI governance.
Penalties for non-compliance can reach up to 35 million euros or 7% of a company’s global annual turnover, whichever is higher. The AI Office will coordinate enforcement with national authorities in each member state, though critics warn that uneven implementation across countries could create regulatory fragmentation within the single market.
The EU AI Act is widely considered the most comprehensive AI regulation in the world, and its enforcement is expected to shape global norms as other jurisdictions develop their own frameworks. With today’s deadline now active, companies operating in Europe face immediate pressure to demonstrate compliance or risk significant financial consequences.
Sources: European Commission AI Act, RAIL Compliance Guide, Holland & Knight Analysis
Author: Technology Desk
discussion