Signal EU AI Act August 2026: Key Compliance Deadlines for High-Risk AI Systems

Summary

Core provisions of the EU AI Act take effect on 2 August 2026, covering Annex III high-risk systems used in biometric identification, critical infrastructure, education, employment, essential services such as credit scoring and insurance, law enforcement, and migration and justice administration. An exception applies to Article 6(1), which does not take effect until August 2027, while phase-in of the broader regulation began in February 2025. Providers must complete conformity assessments verifying that systems meet the Act's safety, transparency, and governance requirements, and must maintain technical documentation describing purpose, design, and performance under Article 11. High-risk systems must be registered in the EU database via the AI Act Service Desk before being placed on the market, and non-EU companies must appoint authorized representatives who retain records for ten years. Under Article 99, penalties for primary violations reach up to €15 million or 3% of global annual turnover, while providing misleading information can draw fines up to €7.5 million or 1% of turnover. The AI Office gains enforcement powers on the same August 2026 date, and the Act applies to US companies whenever their systems are sold to EU customers, affect EU residents, or are embedded in EU products.