On July 20, 2026, the European Commission (EC) published new Guidelines on Transparency of AI-Generated Content (available here) to complement the Code of Practice on Transparency of AI-Generated Content it released on June 10, 2026 (available here). The materials arrive just weeks before the AI Act’s transparency obligations take effect on August 2, 2026, and give businesses clearer direction on how to identify, label, and disclose AI-generated content in practice.
For companies with EU operations – or whose AI system output is used in the EU – the Guidelines and the Code offer practical insight into the EC’s compliance expectations and can help turn the AI Act’s transparency requirements into operational steps.
What the Code of Practice Covers
The Code of Practice was developed to help providers and deployers of generative AI systems comply with the transparency obligations set out in Article 50(2), (4), and (5) of the AI Act. Across 38 pages, it sets out practical measures designed to ensure that AI-generated content can be identified, detected, and appropriately disclosed.
The Code is divided into two sections: one for providers, focused on marking and detection of AI-generated content, and one for deployers, focused on labelling deepfakes and AI-generated or AI-manipulated text. Each section is organized around four core commitments that translate the AI Act’s transparency requirements into practical compliance measures.
AI Providers
For AI providers – broadly, organizations that develop and supply AI systems – the Code focuses on four key areas: marking AI-generated or AI-manipulated content, ensuring those markings are detectable, using appropriate marking and detection solutions, and testing and monitoring transparency compliance.
The Code also provides practical guidance on implementation. It sets expectations for machine-readable marking techniques, including digitally signed metadata, imperceptible watermarking, fingerprinting, and logging mechanisms. It also explains when single-layer or multi-layer marking approaches may be appropriate, when alternative techniques may be used, and which actors may be involved (such as specialized providers of marking and detection technologies).
AI Deployers
For AI deployers – broadly, organizations that use AI systems in their own operations, products, or services – the commitments focus on disclosing deepfakes and AI-generated text, establishing internal compliance procedures, addressing artistic and creative works, and applying human review and editorial control to published content.
The Code also provides practical guidance on these requirements. Notably, it sets design expectations for disclosures indicating that content has been generated or manipulated by AI, including standardized icons and labels (in Annex I). The Code also includes useful guidance for organizations in the media sector, explaining how the AI Act’s transparency obligations interact with existing audiovisual and editorial rules and standards.
What the Guidelines Clarify
Beyond the Code, the Guidelines clarify the AI Act’s transparency framework more broadly. They explain the scope of the transparency obligations, the exceptions that may apply, and how those obligations interact with other AI Act requirements, including rules for high-risk AI systems. They also address individual transparency requirements for interactive AI systems, marking and detecting AI-generated or AI-manipulated content, emotion recognition systems, biometric categorization systems, and labelling deepfakes and AI-generated text. For international businesses, these clarifications are useful because they help identify which technologies and use cases may fall within the scope of EU rules and how the requirements are expected to be enforced in practice.
Why These Tools Matter
The Code serves as a voluntary tool that providers and deployers of AI systems can use to demonstrate compliance with the AI Act’s transparency requirements. Organizations may adhere to the Code by signing it and submitting a signatory form to the EC through the dedicated online portal (available here). Providers and deployers can comply with the transparency requirements through other means – but they will need to be able to demonstrate that such means are adequate.
The Guidelines, by contrast, were developed by the EC with input from EU Member States, the AI Board, and a broad range of stakeholders through a public consultation process. Although they are not legally binding, they provide authoritative insight into how the EC interprets the AI Act’s transparency provisions and expects organizations to implement them.
What Businesses Should Do Now
With the AI Act’s transparency obligations taking effect on August 2, 2026, organizations that sell or use AI systems in the EU (or whose AI system output is used in the EU) should assess now whether their existing practices meet the new requirements. That assessment should cover relevant AI systems, customer-facing content, marketing assets, support tools, and EU operations or services that may involve AI-generated or AI-manipulated content.
The A&B Privacy, Cyber & Data Strategy team is available to help businesses understand these new obligations, assess their business impact, and translate them into practical, operational compliance measures.
