The regulation arrives in stages
The AI Act entered into force on 1 August 2024 and lands in stages. Prohibitions and AI literacy applied from 2 February 2025, GPAI obligations and penalties from 2 August 2025, and the general application — including Article 50's transparency duties — from 2 August 2026. The high-risk deadlines were then pushed back by Regulation (EU) 2026/1744.
Why it is like this
Each stage needs machinery that did not exist when the Act passed — notified bodies, harmonised standards, the AI Office. When the standards slipped, the deadlines that depended on them moved rather than the obligations being watered down.
Not to be confused with
A single deadline, or a settled one. Annex III high risk now applies from 2 December 2027, Annex I product-embedded high risk from 2 August 2028, and high-risk systems used by public authorities have until 2 August 2030. The bans and the literacy duty, meanwhile, have applied since February 2025 — an organisation that has done nothing is already late, not early.
In practice
The deferral did not move everything. Article 50 transparency still applied from 2 August 2026, and two new prohibitions — AI-generated non-consensual intimate imagery and child sexual abuse material — apply from 2 December 2026.
Source: Regulation (EU) 2026/1744 (Digital Omnibus on AI), amending Regulation (EU) 2024/1689 — Article 113 of Regulation (EU) 2024/1689, as amended by Regulation (EU) 2026/1744. Read on .
Odd one out Check yourself
Three of these AI Act obligations apply today. Which one does not?
Show the answer
The Annex III high-risk obligations
Annex III high risk was deferred to 2 December 2027 by Regulation (EU) 2026/1744. The prohibitions and the literacy duty landed in February 2025, and Article 50 transparency on 2 August 2026.