EU AI Act's Core Transparency and GPAI Oversight Provisions Now Enforceable

On August 2, 2026, significant portions of the European Union's Artificial Intelligence Act officially became enforceable, marking a pivotal moment in global AI regulation. Specifically, the transparency obligations outlined in Article 50, which mandate disclosure for AI-generated content like chatbots and deepfakes, are now fully active. Concurrently, the EU AI Office has gained supervisory and fining powers over providers of General Purpose AI (GPAI) models. This immediate applicability comes despite the widely publicized deferral of compliance deadlines for standalone high-risk AI systems, a result of the 'Digital Omnibus' legislative amendments. The distinction is crucial: while some complex requirements have been pushed back, core transparency and foundational model oversight are here now. This development matters immensely to any organization developing, deploying, or utilizing AI systems that operate within or serve the EU market. The era of abstract AI ethics guidelines is over; companies are now facing concrete legal obligations with substantial financial penalties, potentially reaching €15 million or 3% of global annual turnover, whichever is greater. For practitioners, this means a direct and immediate impact on how AI systems are designed, implemented, and monitored. The focus shifts from merely understanding the Act to actively demonstrating compliance, particularly for user-facing AI applications and any services built upon large language models or other GPAI. Failure to comply can lead not only to hefty fines but also reputational damage and a loss of trust among customers and partners. This immediate enforcement aligns with a broader, well-established trend towards increasing regulatory scrutiny of AI worldwide. The EU AI Act has long been recognized as a pioneering, comprehensive framework, and its phased implementation reflects the global challenge of governing rapidly evolving AI technologies. While other jurisdictions, such as the US with its voluntary safety frameworks, are also moving towards AI governance, the EU's approach is notable for its prescriptive nature and robust enforcement mechanisms. The Act's influence extends beyond EU borders, as its categories and requirements often become de facto global standards for companies operating internationally. The current enforcement phase underscores the growing consensus that AI governance is no longer optional but a critical component of responsible technological advancement, moving from theoretical discussions to tangible legal requirements. In practice, practitioners must take immediate steps. First, conduct a thorough inventory of all AI systems, especially those interacting with end-users or relying on GPAI, to identify where Article 50 transparency requirements apply. This includes ensuring clear and accessible disclosure when users are interacting with a chatbot or when synthetic media is presented. Second, if your organization provides GPAI models, understand the new obligations regarding technical documentation, risk assessments, and potential scrutiny from the EU AI Office. Third, establish or update internal AI governance frameworks to explicitly address these new compliance mandates, including clear accountability, audit trails, and incident response plans. The deferral for high-risk systems should not create a false sense of security; the immediate enforcement of transparency and GPAI oversight means that a proactive, rather than reactive, approach to AI governance is paramount to mitigate legal and operational risks.