What the EU's AI labelling code means for your writing
What was announced?
On June 10, 2026, the European Commission published the final version of its voluntary Code of Practice on marking and labelling AI-generated content. The code operationalizes the transparency duties in Article 50 of the AI Act, which apply from August 2, 2026. In broad strokes, it describes machine-readable marking, such as watermarks and metadata, for providers of generative AI systems, and disclosure duties for professional deployers. That includes labelling AI-generated text published to inform the public on matters of public interest.
The code itself is voluntary. The underlying obligations are not. Those come from the AI Act and arrive on schedule.
What does it mean for personal, on-device writing?
The shift is simple. Until this week, “should AI-assisted text be labelled” was a debate. Now there is a written reference point. Providers are expected to mark content at the technical level. Professional deployers are expected to disclose when AI-generated text reaches the public on matters of public interest.
Notice what the framework keeps pointing at: published content, and the role a human played in it. Most everyday writing is not in scope at all. An email to a client, a reply to your landlord, a message to your manager — that is private correspondence, not public-interest publishing.
Where you do publish, the human role is the question that matters. A fully synthetic article posted untouched sits differently than a draft you shaped, corrected, and signed off on. Tools that keep your voice and your final judgment in the loop sit comfortably in this picture. If you publish at scale, or work in journalism, read the code itself rather than summaries of it. It is the document your editors and clients will be quoting from.
Where does ILURA stand?
This is an explainer, not a victory lap. The duty follows what you publish, not which app you used to draft it. ILURA keeps a human firmly in the loop: role playbooks you train by correction, generation on your device, and you decide what ships. None of that exempts anyone from anything. If text you publish falls under Article 50, the same disclosure questions apply to you as to everyone else. For most ILURA users, writing messages rather than publishing articles, the practical change is zero.
Read the signal through ILURA
Platform news matters when it changes what users expect from personal AI. ILURA reads these shifts through one lens: private agents trained by the user on iPhone.
- What becomes possible?
- What should stay user-controlled?
- What belongs on device?
Try it now
Put this to work on a real message.
Open ILURA, bring in a message you actually need to handle today, and get it done in your voice — free, on device, no account. It learns the preference, so the behavior carries to the next one.
Free to start · No account · Data Not CollectedQuick answers
- Does the EU code apply to private emails and messages?
- No. The transparency duties in Article 50 concern content that reaches the public, including AI-generated text that informs the public on matters of public interest. A private email or chat message is not published content.
- Is the Code of Practice legally binding?
- No, it is voluntary. It shows providers and deployers a practical way to meet the AI Act's Article 50 transparency obligations, which apply from August 2, 2026.