Signal EU AI Act High-Risk and Transparency Obligations Become Legally Binding on August 2, 2026

Summary

Substantive requirements for high-risk AI systems under the EU AI Act (Regulation (EU) 2024/1689) became legally binding on August 2, 2026, covering provider obligations under Articles 9-17 (risk management, data governance, technical documentation, record-keeping, transparency, human oversight, and performance standards) and deployer obligations under Article 26. The same date activated Article 50 transparency rules requiring AI chatbots to identify themselves to users and mandating machine-readable watermarks on AI-generated or manipulated content such as deepfakes; the EU also announced a new specialist enforcement team to combat deepfakes and AI-driven disinformation. Providers must complete conformity assessments, issue EU declarations of conformity, affix CE marking, and register systems in the EU database, with non-compliance under Articles 8-15 carrying penalties of up to EUR15 million or 3% of global annual turnover, whichever is higher. Under the European Commission's November 2025 'Digital Omnibus' proposal, Annex III high-risk category obligations may be postponed to December 2027 (stand-alone systems) and August 2028 (product-embedded systems), a delay the European Parliament voted to support on March 26, 2026; however, until that legislation is formally adopted, the original August 2, 2026 date remains binding. Systems placed on the market before August 2, 2026 are generally grandfathered unless significantly modified.