EU Sets Out Practical Path for AI Content Transparency

The European Commission has published a voluntary Code of Practice on marking and labelling AI-generated content, marking a crucial step toward implementing the transparency requirements of the EU AI Act that take effect on August 2, 2026.

Key Developments

The Code of Practice, developed through a multi-stakeholder process facilitated by the AI Office, provides practical guidance for AI providers and deployers to meet upcoming legal obligations. From August 2, 2026, the AI Act will require clear labelling in specific cases: deepfakes and AI-generated or AI-manipulated text published on matters of public interest must be distinctly marked.

The Commission is now inviting all providers and deployers to sign the Code, positioning it as a collaborative framework for industry compliance. The Code will be complemented by official Commission guidelines clarifying the scope of legal obligations and addressing aspects not covered by the voluntary framework.

Industry Context

This development reflects the EU’s commitment to balancing AI innovation with public trust and democratic integrity. Deepfakes and manipulated content pose genuine risks to election integrity, public discourse, and citizen trust—concerns that have intensified as generative AI capabilities have advanced.

The voluntary Code approach represents a middle ground between heavy-handed regulation and light-touch oversight. By setting clear expectations before mandatory requirements kick in, the Commission is giving industry time to develop robust labelling standards and integrate them into production workflows.

Practical Implications for Builders and Users

For AI developers and deployers, the Code signals that labelling infrastructure should become standard practice across generative AI systems. Those who sign the Code gain clarity on compliance expectations and demonstrate commitment to responsible AI practices—valuable positioning as regulators globally increase scrutiny.

For creators and platforms, clear technical standards on what constitutes adequate labelling are essential. The forthcoming Commission guidelines will be critical in translating these requirements into implementable processes.

For users and the public, mandatory transparency from August 2026 onwards should improve ability to identify AI-manipulated content, though the effectiveness will depend on widespread adoption and technical robustness of labelling methods.

Irish and European Implementation

Ireland’s National AI Office, expected to be established by August 2026, will play a coordinating role alongside other Market Surveillance Authorities across the EU. This timing aligns with the Code’s implementation, suggesting Irish regulatory bodies should begin developing expertise in monitoring compliance now.

Open Questions

Key uncertainties remain: How will the Commission assess adequacy of submitted Codes? What happens if adoption rates remain low? How will technical standards handle edge cases like AI-assisted content that blends human and machine creation? Will labelling standards be interoperable across platforms, or will fragmentation occur?

The August 2026 deadline is now just weeks away—implementation success will depend on swift industry action and clear regulatory guidance in the months ahead.

Source: European Commission