top of page
Search

A New Phase for the EU AI Act: European Parliament Approves Legislative Amendments

  • 2 hours ago
  • 2 min read

Selin Çetin Kumkumoğlu

Yaren Alparslan

Buse Sığın

Following the plenary vote held on 16 June 2026, the European Parliament approved, by a large majority, the legislative proposal introduced by the European Commission as part of its digital regulatory package to amend the European Union Artificial Intelligence Act (“EU AI Act”).

While preserving the fundamental structure and risk-based approach of the EU AI Act, the amendments aim to simplify compliance obligations, enhance legal certainty, and reduce the regulatory burden imposed on entities subject to the Act.

Key amendments introduced to the EU AI Act are summarized below:

Postponement of Application Dates for High-Risk AI Systems

The obligations applicable to stand-alone high-risk AI systems listed under Annex III of the EU AI Act will apply from 2 December 2027.

For products incorporating AI components that are regulated under the European Union’s product safety and market surveillance framework, the relevant obligations will apply from 2 August 2028.

Transparency Obligations for AI-Generated Content

The application date of the obligations concerning the detection and labelling of AI-generated content has also been postponed.

Accordingly, the transparency obligations set out in Article 50 of the EU AI Act, which require the disclosure of AI-generated content, will become applicable on 2 December 2026.

Prohibition of “Nudifier” Applications and Systems Generating Child Sexual Abuse Material

The amendments expand the list of prohibited AI practices under the EU AI Act.

In this regard, AI systems capable of generating child sexual abuse material, as well as AI systems capable of creating images, videos, or audio recordings depicting an individual as nude, partially nude, or engaged in sexually explicit conduct without that individual’s consent, are prohibited from being placed on the EU market.

The prohibition applies not only to providers of such systems but also to individuals and organisations that use these systems for the prohibited purposes.

Elimination of Overlapping Obligations Between Product Safety Legislation and the EU AI Act

The amendments introduce measures aimed at reducing overlapping obligations arising from the interaction between the EU AI Act and sector-specific product safety legislation, particularly in relation to machinery products.

Under the revised framework, AI functionalities that merely provide user assistance or optimise product performance without creating health or safety risks will no longer be automatically classified as “high-risk”. In addition, the concept of a “safety component” has been further clarified.

Processing of Personal Data for Bias Detection and Mitigation

The amendments broaden the exceptions allowing the processing of personal data for the purposes of identifying and mitigating bias in AI systems.

Accordingly, where necessary and subject to the implementation of appropriate technical and organisational safeguards, personal data may be processed for these purposes in both high-risk and non-high-risk AI systems.

Conclusion

Companies operating in or targeting the European Union market should not regard the postponement of application dates as a postponement of compliance obligations. Instead, organisations should make effective use of this additional preparation period to align their AI systems with the requirements of the EU AI Act, establish appropriate governance frameworks, and ensure readiness for compliance, particularly with respect to prohibited AI practices and transparency obligations.


bottom of page