Lena explains how rules written in one parliament become the default for the entire internet. A former competition lawyer, she reads the regulations in the original and translates them into plain consequences.
The headline is that the App Store stays under the Digital Markets Act. The binding part is a procedural rule that now tells every designated gatekeeper the same thing: comply first, litigate later.
Europe's top court has closed the last appeal in the Android case. Under a 2014 directive, that quietly converts eight years of findings into something every rival Google shut out can now carry into a national court — without having to prove the case again.
Brussels can already compel Google to hand rivals its query and click data. Whether that data arrives in real time or too late to matter is the whole case — and the deadline is July 27.
The Digital Omnibus pushes the AI Act's high-risk rules to December 2027. The clause worth reading is the one that changes what counts as high-risk at all.
The headline is a ban. The binding text is a delegated power, a list of "reasonable steps" platforms will argue about, and an age-check standard that — like Australia's — every app ends up building once and applying everywhere.
Interim measures — used twice since 2001 — order Meta to reopen WhatsApp to rival AI assistants on last October's terms, years before the investigation will conclude. Here is what the text requires, who it binds, and why the deadline matters more than the fine.
A new taxonomy names 37 manipulative "dark patterns" in AI chatbots. The interesting question is not whether they are sleazy. It is which ones the existing rulebook already binds.
The obligations on foundation-model providers are already in force. The summaries say the hard part was delayed; the binding text says the clock for the labs ran out last August.
A single clause of the EU AI Act binds anyone whose output reaches Europe — and because no one ships a separate chatbot per market, it is quietly becoming the world's labelling rule.
