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

(155) These Guidelines constitute a first interpretation with practical examples of the transparency obligations laid down in Article 50 AI Act. The Commission will review these Guidelines as soon as necessary in view of practical experience gained in the implementation of the transparency obligations and the pace of technological, societal, and regulatory developments in this area. This also includes any relevant experience from market surveillance enforcement actions and interpretations on Article 50 AI Act given by the CJEU. During such a review, the Commission may decide to withdraw or amend these Guidelines. The Commission encourages providers and deployers of AI systems, national market surveillance authorities through the AI Board, the AI Advisory forum, the research community, and civil society organisations to contribute to this process by responding to future calls for public consultation.

(50) The special grandfathering rule in Article 111(2) AI Act applies to high-risk AI systems placed on the market before the date of the application of the high-risk rules only with respect of the compliance of those systems with the requirements and obligations for high-risk AI systems. If the high-risk AI system is also subject to one or more transparency obligations under Article 50 AI Act, this special grandfathering rule does not apply since the application of the different obligations for high-risk and Article 50 transparency are cumulative and the grandfathering is justified and limited only with regard to the compliance with the high-risk provisions.