Guidelines on the implementation of the transparency obligations for certain AI systems under Article 50 of the AI Act — C(2026) 5054 final, Annex

(153) According to Article 113 AI Act, Article 50 AI Act will apply as from 2 August 2026. This requires all in scope AI systems placed on the market or put into service in the Union to be compliant with that provision on that date, regardless of their date of placement on

(47) As the market surveillance authority for AI systems built on GPAI models provided by the same provider (Article 75(1) AI Act). This covers the obligation applicable to providers under Article 50(1) and (2) AI Act. regarding Article 50(3) and (4) AI Act, the AI Office will only be competent if the respective GPAI model and system provider, is also a deployer of the system.

(48) As the market surveillance authority for the EU institutions, agencies and bodies.

(49) Article 85 AI Act.

the market or putting into service (50). Regulation amending the AI Act (the AI Omnibus), which has been recently adopted by the Union legislature, envisages a targeted grandfathering rule only with regard to the marking and detection obligations under Article 50(2) AI Act for generative AI systems placed on the market or put into service before 2 August 2026. It gives providers of those existing systems a transitional period to bring their systems in conformity by 2 December 2026. Systems that are partly interactive and partly generative may benefit from this transitional period only with regard to the marking obligation under Article 50(2) AI Act, while compliance with the disclosure obligation for AI systems directly interacting with natural persons must be ensured as of 2 August 2026.

(154) AI-generated or manipulated outputs falling within the scope of Article 50(2) AI Act and deep fakes within scope of Article 50(4), first subparagraph, AI Act, which have been generated or manipulated before 2 August 2026 do not need to be marked or labelled retroactively. The same applies to texts on matters of public interest falling within scope of Article 50(4), second subparagraph, AI Act that have been AI-generated or manipulated and published before 2 August 2026. Conversely, if texts that have been generated or manipulated before 2 August 2026 are published on or after that date, they need to be labelled. Considering the objective of transparency and increased trust and integrity of the information ecosystem pursued by the AI Act, deployers of AI systems and other actors who are in the possession of or disseminate pre-existing unlabelled deep fakes are, nonetheless, encouraged to do so without being expected to engage in disproportionate efforts to that end (such as auditing of pre-existing content databases or modifying already printed product packaging).