EU AI Act Compliance Timeline: Key Dates and How to Prepare
The EU AI Act entered into force in August 2024 and will apply progressively, with requirements rolling out in phases — step by step. This compliance timeline helps you understand what happens when, and how your organisation can prepare for each phase. From the first bans on prohibited AI practices to the full application of high-risk system requirements, early planning is key to avoiding non-compliance and staying competitive.
Why the Timeline Matters
The EU AI Act doesn’t apply all at once. Its obligations are introduced gradually to give organisations time to adapt. Understanding the timeline is essential for aligning your compliance strategy, budgeting, and training efforts with upcoming regulatory milestones.
Key Milestones at a Glance
This EU AI Act compliance timeline helps organisations understand the key dates and milestones they need to prepare for. In the table below you’ll also find out what each of the AI Act key dates means for your organisation.
EU AI Act Compliance Timeline
| Date | Milestone | What It Means for Organisations |
|---|---|---|
| August 2024 | The AI Act entered into force | The regulation became EU law, starting the countdown for implementation. |
| February 2025 | Prohibited (unacceptable risk) AI systems banned and AI literacy became mandatory | Organisations must ensure they do not use or place banned AI systems on the market, and make sure that their staff have a sufficient level of AI literacy. |
| August 2025 | Requirements for GPAI started to apply | Organisations must ensure that they understand which rules apply when using generative AI. |
| August 2026 | Obligations for high-risk AI systems apply and the main part of the Act | Providers and deployers must meet requirements such as risk management, technical documentation, and conformity assessment. |
| August 2027 | Obligations for high-risk AI systems used as a safety component of a product apply | Product manufacturers must meet requirements such as risk management, technical documentation, and conformity assessment. |
How to Use the Timeline
Each EU AI Act milestone marks a new phase of readiness. To stay on track:
1. Conduct a gap analysis – Compare your current AI governance structure with the Act’s requirements.
2. Set internal deadlines – Start implementation well before official enforcement dates.
3. Assign responsibility – Ensure that legal, technical, and compliance teams work together.
4. Monitor guidance – Follow updates from the European AI Office and national authorities for implementation details.
Prepare Early — and Gain an Advantage
Organisations that start early will not only reduce compliance risk but also gain a competitive edge. Building documentation, establishing governance frameworks, and training teams now will make the transition smoother when the rules become mandatory.
The AI Act isn’t just about regulation — it’s about readiness, trust, and long-term resilience.
Could the Omnibus Package Affects the Timeline?
On 19 November 2025, the European Commission presented the Digital Omnibus Package, which includes proposed amendments to the AI Act aimed at reducing administrative burdens for companies. At this stage, it is impossible to know which parts of the proposal will ultimately be adopted, as the package must undergo negotiations before any decisions are made.
One key element is the suggested postponement of the implementation timeline for the rules governing high-risk AI systems. In the proposed amendments, the Commission aims to link the applicability of the high-risk requirements to the availability of the relevant technical standards.
For high-risk systems under Annex III in the AI Act, the proposal outlines three potential scenarios:
- If the AI Act amendments are not approved and enacted before this date, the current compliance date still applies: 2 August 2026.
- If technical standards are approved on time, the new compliance date would be six months after the approval.
- If the standards are not approved in time, a backstop date of 2 December 2027 would apply.
Checklist to develop a high-level AI compliance roadmap:
- Verify that no systems fall under “unacceptable risk”.
- Create an AI Literacy Training Plan.
- Identify all GPAI or foundation models used or integrated in your organisation.
- Conduct a risk classification of all AI systems based on the Act’s four risk categories.
- Start using an AI Governance or Compliance Tool to track obligations.