AI Policy & Governance, European Policy
CDT Europe’s AI Bulletin: March 2026
March brought major developments in Europe’s AI policy landscape, with policymakers advancing positions on the AI Omnibus, copyright in the age of generative AI, and new rules on AI-generated content. From trilogue preparations to fresh consultations, CDT Europe’s March AI Bulletin keeps you up to speed on the latest EU AI governance developments.
AI Omnibus: European Parliament and Council Finalise their Positions
The European Parliament’s Committee on the Internal Market and Consumer Protection and the Committee on Civil Liberties adopted their joint position on the AI omnibus with a large majority in favour. The compromise text largely upholds key pillars of the Commission’s proposal, and notably expands the scope of the omnibus on several fronts. Among the most significant and controversial changes — aligned with industry demands but highly criticised by a broad range of public interest stakeholders —, it proposes to largely exclude high-risk AI systems that are products or safety components of products regulated by sectoral legislation from the AI Act’s scope, instead leaving it up to said sectoral legislation to cater for the specificities of AI risks. The position also includes a new prohibited AI practice, banning AI systems generating deepfake non-consensual intimate imagery in the absence of effective safety measures being taken by providers and deployers, echoing a similar ban proposed in the Council’s final compromise text.
The Council of the European Union’s compromise position seeks to retain national competence for the oversight of AI systems built on general-purpose AI models where both the system and the model are developed by the same provider or by providers who are part of the same undertaking. In a similar vein, the Council’s compromise text sets out provisions governing the relationship between the AI Office and national authorities.
Both positions reinstate the obligation for providers to register AI systems in an EU-wide database for high-risk AI systems when they consider the system to be exempted from the high-risk classification, and bring back AI literacy obligations for providers and deployers to varying degrees. However, in a blow to a swift application of the AI Act, both the Council and Parliament’s positions uphold the Commission’s proposal to delay the application of the obligations pertaining to high-risk AI systems by over one year. They also support the European Commission’s broadening of the derogation to process sensitive data for bias detection and correction, which now extends to all providers and deployers of AI systems regardless of the risk level that any specific AI system may pose.
The Committee position will now need to be approved in plenary with a vote scheduled for today, 26 March, with trilogue negotiations beginning soon thereafter.
Protecting Copyright in the Age of AI
The European Parliament adopted an own-initiative report to better protect copyrighted work from use by generative AI. MEPs call on the European Commission to ensure that the use of copyrighted material by generative AI is fully transparent and fairly remunerated, to create a new licensing market for copyrighted material and to better protect the press and news media sector from copyright infringements.
The recommendations argue that content fully created by AI should not be copyright protected and highlight the need to protect individuals from the dissemination of manipulated and AI-generated content. Civil society acknowledged several improvements in the final report compared to the original draft, in particular the recognition that the text and data mining exception applies to AI training. However, they argue that the report falls short of offering a viable solution for the licensing market, as it fails to address the structural imbalance in bargaining power that lies at the heart of the problem.
In other news:
- The European Commission published the second draft of the Code of Practice on Marking and Labelling of AI-generated content under the AI Act. The Commission states that this version has been simplified to provide more flexibility for signatories and reduce the compliance burden with several measures being removed or consolidated. The draft is currently open for feedback by participants in the process.
- On 15 March, World Consumer Rights Day, more than 200 organisations and individuals called for an ambitious Digital Fairness Act (DFA) to ensure strong enforcement and protections of consumers’ rights in the digital environment, echoing the EU Council conclusions on the 2030 Consumer Agenda which were similarly positive. Industry on the other hand continues to resist the initiative, arguing that any proposal made prior to the completion of the Digital Fitness Check exercise would be premature.
- The Digital Fitness Check consultation closed on 11 March, showcasing a wide range of responses. Industry responses broadly call for further deregulation beyond the digital omnibus package — which they regard as insufficient — and warn against introducing any new legislation. Civil society organisations underscore existing gaps in ensuring protection for individuals, and stress the need for robust implementation and enforcement of the digital rulebook.
- Following the European Commission’s announcement of possible antitrust interim measures against Meta for blocking third-party AI assistants on Whatsapp, Meta has now moved to allow other AI companies to offer their chatbots on Whatsapp subject to a fee.
- The European Commission opened a consultation on a draft implementing act laying down rules for the enforcement of the AI Act’s rules on general-purpose AI (GPAI). The draft lays out arrangements for the evaluation of GPAI models by the Commission, including the involvement of independent experts, decision-making processes, as well as limitation periods for the imposition of fines. The draft is open for feedback until 9 April.
- The Council of Europe’s Committee of Ministers adopted two recommendations on AI and gender equality. The first recommendation on Equality and Artificial Intelligence aims to assist Member States in promoting equality and preventing discrimination within the lifecycle of AI systems. The second recommendation on Accountability for Technology-Facilitated Violence Against Women and Girls provides guidance on legal, institutional and regulatory measures to prevent and respond to such forms of violence. Council of Europe Recommendations are non-binding.
- An Italian Court overturned the 15 million Euro fine which the Italian Data Protection Authority imposed on OpenAI following an investigation which found several violations of the General Data Protection Regulation (GDPR) by the company’s operation of ChatGPT. At the time of writing, the Court’s reasoning is not yet public.
- Several civil society watchdog organisations called on the coordinators of the European Parliament’s Industry, Research and Energy Committee to withdraw the appointment of MEP Aura Salla as rapporteur for the digital omnibus proposal, highlighting concerns about potential conflicts of interest, given her former role as a Meta lobbyist. The digital omnibus proposal contains AI-specific amendments to the GDPR which lower the bar for data processing.
- French prosecutors alerted the US Department of Justice and the US Securities and Exchange Commission on suspicion that the Grok AI chatbot controversy over sexualised deepfakes was deliberately engineered to artificially increase the value of the company in anticipation of a possible listing on the stock market.
- In an op-ed published by the Financial Times, Mistral CEO Arthur Mensch put forward a proposal for AI developers to pay a content levy of between 1 and 1.5% to flow into a European fund supporting the cultural sector, in exchange for the certainty that AI providers would be shielded from liability for training on materials accessible on the web.
Content of the Month 📚📺🎧
CDT Europe presents our freshly curated recommended reads and works for the month. For more on AI, take a look at CDT’s work.
- Council on Foreign Relations, America’s Digital Empire Has a Trust Problem
- Investigate Europe, Blocked in Europe, deployed abroad: The facial recognition system monitoring Brazil’s schoolchildren
- 7amleh, New Report Examines the Role of EU Funding and Export of High-Risk AI Systems in Escalating Human Rights Violations in Palestine and the Region