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

(89) Article 50(2) AI Act provides three explicit exceptions from the transparency obligations laid down in that provision

(90) The first exception concerns AI systems to the extent that they perform an assistive function for standard editing. Standard editing should be understood as the process of preparing existing content for publication or distribution (e.g., small edits to improve readability and grammar, quality and format) and does not involve generating new content. Standard editing aims at ensuring that content is, among others, free from obvious technical or grammatical linguistic errors, is aligned with applicable lay-out, presentation, formatting or accessibility requirements, or in conformity with sectoral practices. Editing goes beyond standard editing if the content is changed in a material way (substantive modifications, structural changes etc.) that affect its meaning, style or intent.

(91) The second exception applies where an AI system does not substantially alter the input data provided by the deployer or the semantics thereof. An alteration should be considered substantial if the input data or its semantics have been manipulated significantly by the AI system during its output generation process. This must be based on an assessment of relevant factors, such as the format, media content type, style and changes in the content that affect its meaning, style or intent. Whether that is the case requires a case-specific assessment.

(92) If an AI system can be used both for content generation or manipulation and for non- substantial minor alterations of input data, then the transparency obligation under Article 50(2) AI Act will not apply to the content altered in a minor non-substantial manner. Examples of standard editing and minor alterations benefitting from the exception under Article 50(2) AI Act:

- Grammar correction and spellchecking, linguistic and minor stylistic polishing that do not change the substance, meaning, style or messaging of text, AI-generated translations of text;

- Formatting, format conversions, technical compression, noise reduction or removal for enhanced clarity without changing the meaning or the substance of the content;

- Minor cropping, minor colour adjustments or corrections, lightening or darkening, sharpening for enhanced clarity or other standard editing and technical corrections;

- Removal of dust spots caused by a dirty lens or sensor, removal of red-eye caused by flash photography; deleting and obscuring backgrounds that are visible in the original file, pixelation or blurring of faces;

- Rescaling of a video clip, dynamic range compression and equalisation; limited video stabilisation;

- Standard adjustments to colour and contrast, minor adjustments to playback speed, minor corrections to level the horizon of an image, applying pixel filters to amplify certain parts of an image, or applying colour maps to grayscale images, edge image completion, converting a black & white to colour image or video and vice versus;

- Pixel filling for aspect ratio adaptation or enhanced clarity, automatic transition clips, other non-substantive edits for cosmetic and technical purposes.

- Medical image technical processing or reconstruction within medical devices (e.g. annotations, boxes, contours, circles, heatmaps);

- Transcriptions of conversations; AI-generated content that only transforms authentic human input through assistive technologies allowing persons with disabilities to communicate (e.g., augmentative and alternative communication (AAC) or customized neural voices (CNV)) since they do not alter semantically the meaning of the content. Examples of semantic changes that require marking under Article 50(2) AI Act:

- AI-generated summaries of text; paraphrasing or rewriting text that changes style, structure and meaning beyond mere grammatical and minor stylistic correction;

- Removal, replacement or insertion of objects or persons in existing images and videos that changes meaning and substance of the content; face replacement or substantial facial modification,

- Synthesis of realistic speech in a specific person's voice, or generation of realistic video depicting events that did not occur; altering the body shape or the skin colour of a person; extreme lightening, darkening, colour and contrast changes that change the meaning, intent and messaging of the content;

- Creation of composite images or video clips that modifies the representation of persons, objects, events or facts, any other substantial alteration of the content.

(93) The third exception applies to generative AI system authorised by law to generate or manipulate synthetic content to detect, prevent, investigate or prosecute criminal offence The clarifications provided in points 46-48 above for this exception are also relevant in this case.