Skip to main content

02 · Implementation Timeline

Sources

Phased application​

The Act entered into force 1 August 2024 but applies progressively.

Wave 1 — 2 February 2025 ✅ (past)​

  • Prohibited AI practices (Article 5) — 8 banned categories
  • AI literacy obligations for all providers and deployers
  • Chapter I (general provisions) and Chapter II apply

Wave 2 — 2 August 2025 ✅ (past)​

  • GPAI model obligations (Chapter V)
  • Governance framework (Chapter VII): AI Board, Scientific Panel, Advisory Forum must be set up
  • Notified bodies chapter (Chapter III, Section 4)
  • Penalties articles (99 and 100)
  • Member states must designate national competent authorities (notifying authorities + market surveillance authorities)
  • Member states must adopt national penalty laws
  • GPAI models placed on market before this date get a 2-year compliance grace period (until 2 Aug 2027)

Intervening milestone — 27 July 2026 ✅ (past)​

  • Regulation (EU) 2026/1744 (Digital Omnibus on AI) entered into force
  • Article 4 now requires providers and deployers to take measures that support the development of AI literacy
  • New Article 4a creates an exceptional, safeguard-bound legal basis for processing special-category personal data when strictly necessary for bias detection and correction
  • AI Office supervision and enforcement powers expanded for specified AI systems under Article 75
  • Articles 102–110 became applicable

Wave 3: 2 August 2026 (past)​

  • Base AI Act: general application date for remaining provisions not otherwise delayed.
  • Article 50 transparency obligations have applied since this date. Regulation (EU) 2026/1744 gives providers of synthetic-content systems placed on the market before 2 August 2026 until 2 December 2026 to comply with Article 50(2).
  • The national AI regulatory sandbox deadline is 2 August 2027.
  • Chapter III, Sections 1–3 do not apply to Annex III systems on this date. Regulation (EU) 2026/1744 moves them to 2 December 2027.

Wave 4 — 2 August 2027​

  • Providers of GPAI models placed on the market before 2 August 2025 must comply by this date.
  • National AI regulatory sandboxes must be operational by this date.
  • Large-scale IT systems (Annex X) placed on the market or put into service before this date must comply by 31 December 2030.

Wave 5 — 2 December 2027​

  • Chapter III, Sections 1–3 apply to AI systems classified as high-risk under Article 6(2) and Annex III.

Wave 6 — 2 August 2028​

  • Chapter III, Sections 1–3 apply to AI systems classified as high-risk under Article 6(1) and Annex I.
  • Commission evaluates AI Office functioning.

Further out​

  • 2 Aug 2030: public-authority deployers of high-risk AI must comply
  • 31 Dec 2030: large-scale IT systems (Annex X) compliance deadline
  • 2 Aug 2031: Commission enforcement assessment report

Key article references​

MilestoneArticle
Entry into forceArt. 113
Prohibited practicesArt. 5, Art. 113(a)
GPAI obligationsChapter V (Arts. 51–56), Art. 113(b)
Notified bodiesChapter III, Section 4
High-risk (Annex III)Art. 6, Annex III
High-risk (Annex I products)Art. 6(1), Annex I
TransparencyArt. 50
GovernanceChapter VII
PenaltiesArts. 99–100
Transitional provisionsArt. 111

Commission milestones (obligations on the Commission)​

  • 2 May 2025: codes of practice shall be ready (Art. 56(9))
  • 2 Feb 2026: guidelines on Article 6 (high-risk classification) to be published (Art. 6(5))
  • 2 Aug 2026 (and annually): Commission reviews and may amend the prohibited practices list (Art. 112(1))
  • 2 Aug 2028 (and every 4 years): review of Annex III use cases, Article 50, governance system (Art. 112(2))
  • 2 Aug 2029: Commission evaluation and report (Art. 112(3))
  • Delegated acts power expires 1 Aug 2029 unless extended (Art. 97(2))

The Commission published final Article 50 transparency guidelines on 20 July 2026. The high-risk classification guidelines remained in draft when rechecked on 7 September 2026, after a targeted consultation closed on 23 July.