Text · Amendment list
Simplification of the implementation of harmonised rules on artificial intelligence (Digital Omnibus on AI)
Report A-10-2026-0073-AM-097-102 · COM(2025)0836 – C10-0304/2025 – 2025/0359(COD)
- Kind
- Amendment list A-10-2026-0073-AM-097-102
- Date
- 23 March 2026
- Dossier
- 2025/0359(COD)
More facts (2)
- Formats
- Official page PDF Word
- Reference
- COM(2025)0836 – C10-0304/2025 – 2025/0359(COD)
Text
The text as parsed from the official Word file. Every paragraph has a link (¶) and can be saved to a project as a passage.
Jump to an amendment (6)
| Text proposed by the Commission | Amendment |
|---|---|
| The authorities involved in the application of this Regulation shall cooperate actively in the exercise of these powers, in particular where enforcement actions need to be taken in the territory of a Member State.; | The authorities involved in the application of this Regulation shall cooperate actively in the exercise of this Regulation.; |
Simplification of the implementation of harmonised rules on artificial intelligence (Digital Omnibus on AI)
| Text proposed by the Commission | Amendment |
|---|---|
| (26a) in Article 78, the following paragraph 6 is inserted: | |
| '6. Where, for the purposes of this Regulation, a competent authority or the Commission requires access to the source code, the provider may request that the inspection, audit or verification shall be conducted on-site at the provider's premises or within a secure, air-gapped digital environment, under the provider's technical supervision. | |
| This request shall be granted where the provider demonstrates that the documentation or code is critical to the cybersecurity, integrity, or safe operation of the provider’s AI systems, or of third-party AI systems that depend on said source code. Such an arrangement shall not limit the authority's power to conduct a full inspection, audit or verification as foreseen under this Regulation, but shall govern the physical and digital custody of the materials. |
The AI source code represents the 'master key' to their logic. If this code is extracted from the provider’s secure environment and stored in lower-security government or agency databases, it becomes a high-value target for state-sponsored actors and cyber-criminals. A leak of such code could allow for the development of 'adversarial attacks', leading e.g. to exposure of critical infrastructure. This systemic risk is already recognized in Union law, specifically under Article 30(3) of Regulation (EU) 2022/2554.
Simplification of the implementation of harmonised rules on artificial intelligence (Digital Omnibus on AI)
| Text proposed by the Commission | Amendment |
|---|---|
| (26b) The following article is inserted: | |
| 'Article 88a | |
| The AI Office shall maintain records of communication with providers and deployers of general-purpose AI models with systemic risk and providers of AI systems falling under the supervision of the Commission under Article 75 and under Article 88 of this Regulation. Such records shall be kept in a consistent manner and made available, where appropriate, in accordance with Regulation (EC) 1049/2001 or other applicable Union or national law and without prejudice to the protection of trade secrets, for the purposes of parliamentary oversight, cooperation between authorities and judicial review.’ |
Establishing a clear obligation to record and systematically organise supervisory communications is necessary to ensure transparency, accountability, democratic oversight and equal treatment of economic operators, in particular with regard to the interpretation and application of the notion of systemic risk and related obligations. Furthermore, it would enable effective judicial and parliamentary scrutiny, while fully preserving the protection of trade secrets.
Simplification of the implementation of harmonised rules on artificial intelligence (Digital Omnibus on AI)
| Text proposed by the Commission | Amendment |
|---|---|
| 2. Without prejudice to the application of Article 5 as referred to in Article 113(3), third paragraph, point (a), this Regulation shall apply to operators of high-risk AI systems, other than the systems referred to in paragraph 1 of this Article, that have been placed on the market or put into service before the date of application of Chapter III and corresponding obligations referred to in Article 113, only if, as from that date, those systems are subject to significant changes in their designs. In any case, the providers and deployers of high-risk AI systems intended to be used by public authorities shall take the necessary steps to comply with the requirements and obligations laid down in this Regulation by 2 August 2030.; | 2. Without prejudice to the application of Article 5 as referred to in Article 113(3), third paragraph, point (a), this Regulation shall apply to operators of high-risk AI systems, other than the systems referred to in paragraph 1 of this Article, that have been placed on the market or put into service before the date of application of Chapter III and corresponding obligations referred to in Article 113, only if, as from that date, those systems are subject to substantial modification in their designs. In any case, the providers and deployers of high-risk AI systems intended to be used by public authorities shall take the necessary steps to comply with the requirements and obligations laid down in this Regulation by 2 August 2030.; |
To align the paragraph with the 2022 NLF Blue Guide standard term that is also used in the remaining parts of the EU AI Act.
Simplification of the implementation of harmonised rules on artificial intelligence (Digital Omnibus on AI)
| Text proposed by the Commission | Amendment |
|---|---|
| 4. Providers of AI systems, including general-purpose AI systems, generating synthetic audio, image, video or text content, that have been placed on the market before 2 August 2026 shall take the necessary steps in order to comply with Article 50(2) by 2 February 2027.; | 4. Providers of AI systems, including general-purpose AI systems, generating synthetic audio, image, video or text content shall take the necessary steps in order to comply with Article 50(2) by 2 August 2027.; |
The Commission proposal was already insufficiently adapted to the practical realities of compliance. The Committee compromise backtracks even further by accelerating the timeline, disregarding the complexity of implementation and the need for realistic preparation.
Simplification of the implementation of harmonised rules on artificial intelligence (Digital Omnibus on AI)
| Text proposed by the Commission | Amendment |
|---|---|
| (d) Chapter III, Sections 1, 2, and 3, shall apply following the adoption of a decision of the Commission confirming that adequate measures in support of compliance with Chapter III are available, from the following dates: | (d) Chapter III, Sections 1, 2, and 3, with the exception of Article 6(5), shall apply on 2 August 2028. |
| (i) 6 months after the adoption of that decision as regards AI systems classified as high-risk pursuant to Article 6() and Annex III, and | |
| (ii) 1 months after the adoption of the decision as regards AI systems classified as high-risk pursuant to Article 6(1) and Annex I. | |
| In the absence of the adoption of the decision within the meaning of subparagraph 1, or where the dates below are earlier than those that follow the adoption of that decision, Chapter III, Sections 1, , and 3, shall apply:(i) on December 07 as regards AI systems classified as high-risk pursuant to Article 6() and Annex III, and (ii) on August 08 as regards AI systems classified as high-risk pursuant to Article 6(1) and Annex I.; |
In order to simplify the provisions on the date of application, a single harmonised date in 2028 is introduced in place of the complicated and conditional mechanism proposed by the Commission. This also better reflects the realities on the ground, notably the complexity of the legal framework, the time needed for the development of standards, and the practical challenges businesses face in preparing for compliance.
Connections
The dossier, the decisions on this text and its other versions.
No connections found for this item.
Sources & citation
Where the facts on this page come from, and how to cite it.
- Data source
- Licensed CC BY 4.0.
- Retrieved
- 25 September 2026
Cite as
European Parliament (2026). “A10-0073/97 - Amendement 97 - Piotr Müller, Assita Kanko, Stefano Cavedagna, Kosma Złotowski - au nom du groupe ECR - Rapport A10-0073/2026 - Arba Kokalari, Michael McNamara - Simplification de la mise en œuvre des règles harmonisées concernant l’intelligence artificielle (train de mesures omnibus numérique sur l’IA) - (COM(2025)0836 - C10-0304/2025 - 2025/0359(COD))”. Text, 23 March 2026. docId A-10-2026-0073-AM-097-102, reference A10-0073/2026. EU Parl Watch Research. https://news.eu-parl.st-solutions.dev/texts/A-10-2026-0073-AM-097-102 (retrieved 25 September 2026). Data: EP Open Data API: document record, https://data.europarl.europa.eu/api/v2/documents/A-10-2026-0073-AM-097-102 (CC BY 4.0).
BibTeX
@misc{epw-text-a-10-2026-0073-am-097-102,
author = {{European Parliament}},
title = {{A10-0073/97 - Amendement 97 - Piotr Müller, Assita Kanko, Stefano Cavedagna, Kosma Złotowski - au nom du groupe ECR - Rapport A10-0073/2026 - Arba Kokalari, Michael McNamara - Simplification de la mise en œuvre des règles harmonisées concernant l’intelligence artificielle (train de mesures omnibus numérique sur l’IA) - (COM(2025)0836 - C10-0304/2025 - 2025/0359(COD))}},
year = {2026},
date = {2026-03-23},
howpublished = {\url{https://news.eu-parl.st-solutions.dev/texts/A-10-2026-0073-AM-097-102}},
url = {https://news.eu-parl.st-solutions.dev/texts/A-10-2026-0073-AM-097-102},
urldate = {2026-09-25},
publisher = {EU Parl Watch Research},
note = {Text. docId A-10-2026-0073-AM-097-102, reference A10-0073/2026. Data: EP Open Data API: document record (CC BY 4.0)}
}