Confession: I didn’t come to Article 22 the way I came to the rest of the DSA. Articles 34 and 35 got the dinner-party monologues. Article 28(4) got a formal submission with my name on it. Article 22 just sat there, twelve lines long, sounding like the most boring provision in the whole regulation — a fast lane for flagging illegal content, certified by a regulator, nothing to see here. I almost didn’t look twice.
Then I looked twice, and it turned out to be the one that unsettled me most.
Here’s the pitch behind Article 22, and it’s a good one: platforms get flooded with low-quality user reports, so let specialist organisations — the ones who actually know what illegal hate speech or terrorist content looks like — earn a “trusted flagger” badge, and their notices jump the queue. A child-safety hotline. A hate-speech monitor. People who do this for a living, not randoms mass-reporting things they disagree with. I still think that’s a genuinely good idea. This isn’t an article about tearing the mechanism down.
It’s an article about the one word the law forgot to define properly: independent.
Independent from whom, exactly?
To get trusted-flagger status, an applicant has to show three things: real expertise in spotting the content in question, that it’s diligent and accurate, and that it’s “independent of any provider of online platforms.”
Read that clause again. Independent of platforms. Not independent of governments. Not independent of political parties. Not independent of whoever happens to be writing the cheques. The law spent all its worry on the one relationship — flagger and platform — and left the door wide open on every other relationship that could just as easily compromise a flagger’s judgment about what counts as illegal content.
Once you’re certified, the annual check-in asks you to explain how you’re staying independent — of platforms, again. Nobody asks who funds you. Nobody asks whether your board answers to a ministry. Nobody asks whether the person deciding what gets fast-tracked for removal also holds a public political position on the exact topic they’re flagging.
I want to be careful here, because this is the part where it would be easy to turn a structural argument into a finger-pointing exercise, and that’s not what I’m doing.
Two examples, and one important caveat
In Germany, a trusted flagger called REspect! was certified in 2024. Documents obtained by a member of the European Parliament reportedly show the organisation gets roughly 95% of its funding from a single government ministry. If that figure is accurate, it doesn’t prove anyone did anything wrong. It proves that an organisation almost entirely dependent on one ministry’s money has a fast lane to flag content across every major platform in the EU — and nothing in the DSA required that funding relationship to be disclosed as part of the public record.
In Ireland, the regulator certified the Jewish Representative Council of Ireland as a trusted flagger for antisemitic content in June 2026 — a legitimate and important category of harm, and I want to say clearly that nothing here is a challenge to the fight against antisemitism, which this organisation shares without reservation. An Irish outlet separately reported that the Council’s chair holds a public political position on Ireland’s recognition of Palestinian statehood, and reported that he helped arrange a meeting at the Taoiseach’s office where the trusted-flagger application came up. The same reporting says the regulator declined freedom-of-information requests about that meeting.
I haven’t independently verified either of those reports, and I’m not asking you to treat them as proven. I’m using them for a narrower point: under the current rules, a journalist or a rival civil-society organisation with concerns like these has no statutory tool to test them. The only lever available — a freedom-of-information request — reportedly got refused. That’s not a scandal about one flagger. That’s a design flaw that would produce the exact same silence no matter which community’s flagger someone had questions about.
The warning sitting just outside the DSA’s door
You don’t have to imagine what happens when a content-decision fast lane runs without a transparent, independent standard. The UK’s Online Safety Act isn’t part of the DSA, but it’s a preview of the same failure mode: in 2023, English and Welsh police made over twelve thousand arrests under older communications offences — about thirty-three a day — and fewer than one in ten led to a conviction. That’s not an argument against enforcement existing. It’s an argument that any mechanism moving this fast needs guardrails proportionate to its speed, or it gets used for a lot more than it was built for.
The gap that should bother you regardless of what you believe
Here’s the part I actually can’t look away from. Nearly four in ten Muslims in the EU report having experienced discrimination. Europe’s own hate-speech monitoring found Islam among the most targeted topics online, right behind Judaism. The European Commission’s own coordination process on anti-Muslim hatred describes things getting worse, not better, since October 2023.
And right now, as far as the Commission’s own public register shows, there isn’t a single Muslim-led or Islamophobia-focused organisation holding trusted-flagger status anywhere in the European Union. Not one. Meanwhile organisations focused on antisemitism — rightly — are cited by the Commission itself as the model example of the kind of specialist group Article 22 was built for.
I don’t think that gap exists because any regulator decided Muslim communities don’t deserve a fast lane. I think it exists because the organisations who could hold that role are, on the whole, smaller, younger, and less plugged into national regulatory relationships — which is precisely the capacity gap EqualFaith Europe exists to help close. But “nobody meant it” and “it’s still true” can both be correct at once, and only the second one matters to the child getting harassed online for wearing a hijab, with no dedicated flagger anywhere built to catch it faster.
So you end up with a strange double bind: the community with some of the clearest documented online-hate data in Europe has no fast lane of its own, and at the same time has no real tool to check whether other communities’ fast lanes are being run with the neutrality the law assumes but never verifies. Fixing the first problem and fixing the second turn out to be the same piece of work.
What would actually fix this
None of this needs Brussels to reopen the DSA. Three changes, all inside powers the Commission and national regulators already hold: require flaggers to disclose funding sources above a materiality threshold, and any political affiliations relevant to what they flag — the same basic transparency lobbying rules already demand elsewhere in EU law. Set up an independent, cross-community review panel that can actually look at a contested designation instead of leaving it entirely to the regulator that granted it. And have national regulators publish, alongside their flagger lists, an honest accounting of which communities are and aren’t represented — then actively help under-represented groups build the capacity to apply.
A very familiar shape, once more
If you’ve read anything else I’ve written, you’ll recognise the pattern by now, because it’s always the same one. A right that sounds airtight in the text. A scope clause, a definition, a missing disclosure requirement, quietly doing the work of excusing exactly the failure mode you’d most want covered. The DSA gave platforms a duty to catch systemic risk and then made it hard for researchers to check whether they actually did. It gave children a safety net and forgot to ask which children were falling through it. And now it gave civil society a fast lane for flagging hate — and forgot to ask who’s allowed to drive.
The mechanism is good. The word “independent” just needs to mean what it says, for every community it’s supposed to protect — not only the ones who got there first.