High-risk obligations for stand-alone Annex III systems (Chapter III, Sections 1-3)
The AI Act's general application date in Article 113, second paragraph: 'It shall apply from 2 August 2026.' Article 6(2)/Annex III systems fell under it.
High-risk obligations for AI embedded in Annex I products (Art. 6(1))
Article 113, third paragraph, point (c) as adopted: 'Article 6(1) and the corresponding obligations in this Regulation shall apply from 2 August 2027.' Note that this track never shared the 2 August 2026 date — the amending act replaced point (c) in full.
Article 50(2) machine-readable marking, for generative systems already on the market before 2 August 2026
Article 50(2) applied from the general date; the amending act adds a four-month transition for systems already placed on the market. New systems from 2 August 2026 do not get it.
Two new prohibited practices: AI generating non-consensual intimate imagery, and material within the meaning of Art. 2(c) and (e) of Directive 2011/93/EU
Article 5 as adopted contained no such prohibition. Article 5(1a) narrows the scope; Article 5(1b) excludes manipulation that does not increase exposure.
The AI literacy duty (Article 4) — date unchanged, substance changed
Article 4 was replaced in full. The new text states expressly that the obligation 'does not require providers or deployers to guarantee any specific level of AI literacy of any individual.' A timeline that only tracks dates misses this one entirely.