EU delays AI Act high-risk rules; transparency stays on time
What was announced?
On May 7, 2026, Council and Parliament negotiators reached a provisional agreement on the EU’s Digital Omnibus package, which amends the AI Act. Law firm White & Case summarized the deal. High-risk obligations under Annex III move to December 2, 2027. High-risk obligations under Annex I move to August 2028. New prohibitions are added under Article 5.
One thing does not move. Article 50 transparency obligations — telling people when they interact with an AI system, and marking certain AI-generated content — still apply from August 2, 2026. The agreement is provisional and still needs formal adoption.
What does it mean for personal, on-device writing?
Headlines compress this to “EU delays AI rules.” For people who write with AI, that compression hides the part that matters.
The delayed provisions cover high-risk systems: hiring tools, credit scoring, biometric identification, critical infrastructure. None of that describes a person drafting an email. The provision that does touch everyday AI use is Article 50, the transparency layer. It lands on schedule this August. Disclosure duties for AI systems and labelling duties for synthetic content did not get a reprieve.
So the practical reading is the opposite of the headline. If you write with AI, nothing relevant to you was delayed. The transparency regime arrives within months, exactly as planned.
Where does ILURA stand?
Nothing ILURA does is remotely high-risk under the Act. Helping you draft a follow-up email on your own phone is about as far from Annex III as software gets. The postponement changes nothing for ILURA or its users.
The honest take: this deal buys enterprises time. It does not buy writers anything, and it does not take anything from them either.
Where the Act’s spirit does intersect with ILURA is responsibility. ILURA treats AI as a drafting layer under your control: you pick the role playbook, you correct the output, you send the final text. And because everything runs on-device through Apple Intelligence, with no account and a Data Not Collected privacy label, there is no server-side trail of your writing for any current or future rule to worry about.
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
- Did the EU delay the entire AI Act?
- No. The provisional Digital Omnibus agreement postpones high-risk obligations, with Annex III moving to December 2, 2027 and Annex I to August 2028. Article 50 transparency duties still apply from August 2, 2026.
- Does the AI Act delay affect ILURA?
- Practically, no. ILURA is a personal writing tool that drafts emails and messages on your own iPhone. That is far from the Act's high-risk categories, so the postponement changes nothing for ILURA users.