Home / Reports / EU AI Act Goes Live

Thank you for reading this post, don't forget to subscribe!

Policy Report · The AI Index

The EU AI Act Goes Fully Live

On August 2, 2026, the bulk of the EU AI Act becomes applicable — including high-risk-system obligations and new transparency rules. Here is the timeline, what actually changes, and the May 2026 amendments that moved some deadlines.

By The AI Index · Updated Jun 18, 2026 · 5 min read

The staged timeline

The AI Act was never a single switch. It entered into force on August 1, 2024 and rolls out in phases. The ban on “unacceptable risk” practices and AI-literacy obligations applied first, from February 2, 2025. Rules for general-purpose AI models and the Act’s governance structure followed on August 2, 2025. The big one is August 2, 2026, when the core obligations for high-risk systems and the transparency rules become applicable across the Union.

What actually changes on August 2, 2026

Two things matter most. First, high-risk AI systems listed in Annex III — covering uses such as certain hiring, credit, and essential-services decisions — face the full obligation set: risk management, data governance, technical documentation, record-keeping, human oversight, accuracy, robustness, cybersecurity, plus deployer duties, conformity assessment, post-market monitoring, and incident reporting.

Second, the Article 50 transparency obligations bite: users must be told when they are interacting with an AI system, AI-generated or manipulated content must be labeled, and deepfakes must be identified as such. These are the rules most likely to be visible to ordinary users.

The May 2026 amendments

On May 7, 2026, EU legislative bodies reached a political agreement on amendments — part of a broader “Omnibus” simplification package. The agreement clarifies existing requirements, extends certain compliance deadlines for high-risk systems, and introduces new rules targeting AI-generated intimate content. Notably, enforcement of some AI-content-labeling obligations was deferred — reporting points to a roughly four-month deferral to December 2, 2026. Companies should track the final adopted text, since deadlines have been a moving target.

“The AI Act was never a single switch — it rolls out in phases, and August 2, 2026 is the one that bites.”

Enforcement and penalties

Enforcement is a hybrid model shared between member-state authorities and the European Commission, supported by the European AI Board and an expert scientific panel. Penalties are tiered: the most serious — breaching the prohibited-practices rules — can reach up to €35 million or 7% of global annual turnover, whichever is higher. Each member state must also establish at least one AI regulatory sandbox by August 2, 2026 to help providers test systems under supervision.

Editorial note: Dates and scope reflect the European Commission’s published timeline and reporting on the May 2026 Omnibus agreement. Some deadlines for embedded high-risk systems extend into 2027, and amendment details may shift until the final text is adopted. This is a summary, not legal advice.

Frequently asked

When does the EU AI Act become fully applicable?

The Act entered into force on August 1, 2024 and applies in phases. The largest set of obligations, including high-risk-system requirements and Article 50 transparency rules, becomes applicable on August 2, 2026. Some embedded high-risk uses have later deadlines into 2027.

What are the transparency obligations?

Under Article 50, users must be informed when interacting with an AI system, AI-generated or manipulated content must be labeled, and deepfakes must be identified. These rules become applicable in August 2026, with some content-labeling enforcement deferred to December 2, 2026.

What are the penalties under the EU AI Act?

Penalties are tiered. The most serious breaches, involving prohibited AI practices, can reach up to €35 million or 7% of global annual turnover, whichever is higher. Enforcement is shared between EU member states and the European Commission.

Related: All reports · The Dark Side of AI · AI dark-side statistics