The Act applies in stages. Build a system-level calendar because different rules can apply to different roles and use cases on different dates.
1 August 2024 — entry into force
The Artificial Intelligence Act entered into force. This started the staged application timetable and gave organisations time to identify systems, roles and future duties.
2 February 2025 — prohibitions and AI literacy
The original prohibited-practice rules, relevant definitions and Article 4 AI literacy duties started to apply. Organisations should already have screened uses against Article 5 and taken context-specific measures to support staff AI literacy.
2 August 2025 — governance and GPAI obligations
Governance provisions and obligations for providers of general-purpose AI models became applicable. Providers of models placed on the market before that date have a later transition date under the applicable rules.
2 August 2026 — general application and transparency
Most remaining provisions became applicable, subject to exceptions. Article 50 transparency requirements began applying, and the AI Office and national authorities began exercising the relevant implementation, supervision and enforcement powers.
A limited transition until 2 December 2026 applies to the machine-readable marking and detection duty for certain systems placed on the market before 2 August 2026. This is not a general grace period for Article 50.
2 December 2026 — additional prohibitions and limited transition end
Additional prohibitions concerning systems that generate or manipulate non-consensual intimate material or child sexual abuse material apply. The limited Article 50 marking transition described above also ends.
2027 and 2028 — high-risk system dates
Following the 2026 Digital Omnibus amendments, rules for the Annex III high-risk use cases apply from 2 December 2027. Rules for high-risk systems that are products or safety components covered through the Annex I product route apply from 2 August 2028.
- 2 December 2027: specified high-risk uses such as certain biometric, education, employment, essential-services, law-enforcement, migration and justice uses.
- 2 August 2028: high-risk AI embedded in or acting as regulated products under the Annex I route.
Turn the legal timeline into a delivery plan
For every system, map the relevant rule, role, application date, control owner, evidence needed and internal delivery date. Work backwards for procurement, technical changes, conformity work, training and governance approval.
- Track official amendments, guidelines, standards and national authority information.
- Do not wait for the final date to gather technical and operational evidence.
- Review the calendar whenever classification or intended purpose changes.
Dates can depend on the exact provision and transition rule. Verify the current official text for your system before relying on a deadline.
Official sources
Use the official text and current Commission guidance for decisions about a specific system.