X has announced that it will start telling you when a government has asked for the visibility of your posts to be reduced. The feature lives in a profile section called “under the hood”, and it is meant to show whether your reach has been cut to comply with local law, naming the country that filed the restriction request and listing the specific messages affected. The stated purpose is to put a formal frame around content filtering and to take the heat out of the accusations about shadowbanning, the practice of quietly throttling someone’s reach without telling them.
Credit where it is due, and it is due. “Am I being suppressed?” is the kind of question that gets treated as an embarrassing hobby of people who simply were not interesting enough, and the user asking it is often left with amateur augury: staring at impression counts, comparing them with last month, building private theories out of noise. A system that replaces that with a named country and a list of affected posts is a genuine improvement. It also moves the ground under the argument, because a company does not build an interface for disclosing restriction requests unless there are restriction requests to disclose.
The eligibility criteria are the story
Then you read the access rules, and the applause dies in your throat.
Consultation is not general. It is reserved, for now, to a restricted circle, and the published criteria are specific: accounts continuously active for more than a year, with a minimum of ten posts published in the previous calendar month. I am not going to invent a rationale on the company’s behalf, and I am not claiming there cannot be one. What I can describe is how those numbers read from outside, and they read like engagement metrics, the same kind of figure a loyalty programme uses to decide whether you qualify for the lounge. A phased rollout can be defended. A phase gated by how often you post is a criterion that tells you something about the gate, and it deserves to be read closely rather than waved through.
So the right to know that a state has moved against your speech arrives, for now, attached to a posting record. Quiet September? Transparency will see you next month. Joined in March? Come back when you have accrued seniority. The information concerns an act of public authority exercised against a private citizen, and access to it currently sits behind activity thresholds, administered by the very company that carried out the restriction on that authority’s behalf.
There is a second-order effect worth naming. If knowing whether you were restricted depends on maintaining a posting cadence, the system rewards the behaviour it measures. And some of the people with the sharpest reason to look, new accounts, occasional posters, somebody who wrote one inconvenient sentence last spring, fall outside the criteria as published.
A receipt is not a remedy
The deeper problem sits underneath the eligibility question, and it would survive even if the criteria were dropped tomorrow.
Being told about a restriction is not the same as having it lifted. The reduced visibility stays reduced. The request stays executed. The government stays satisfied. What changes is that you now receive documentation. That is worth having, and it is not a small thing: documentation is what makes complaint, litigation, journalism and public pressure possible at all. But let us be precise about what this is, which is a disclosure surface. Whatever recourse exists elsewhere, in law or in the platform’s own rules, a panel that shows you a country’s name is not itself a route back to the reach that country took away. A dashboard is not due process. It is the digital equivalent of a “we apologise for the inconvenience” sign taped to a shutter that is already down.
It is also worth keeping the categories straight. A local restriction is not the same as deletion, and a reduced reach is not the same as an account wiped off the platform. The post may still exist, still be readable, still be reachable by someone who goes looking for it. That distinction matters, and flattening it would do the argument no favours. The point is narrower and harder to wriggle out of: a specific state asked for a specific piece of speech to be seen by fewer people, the platform complied, and whether you are allowed to see that record depends, for now, on your posting history.
This expands an earlier test run in August, built to explain the recommendation algorithm’s criteria and to display the moderation labels applied to posts. That origin is informative, not because it reveals anybody’s private intentions, but because it tells you what kind of surface this is: a product-explanation panel, the sort of thing that reassures users about ranking. Government restriction requests sit badly in that frame. They belong to the category of state action against speech, where the questions are legal ones and the answers are owed to everybody, not to a qualifying tier.
What would actually count
Here is the yardstick I would use, and anyone is welcome to measure the finished product against it. Disclosure available to every account from day one, with no activity threshold. Aggregate numbers out in the open: how many requests, from which states, how many complied with, how many refused. A stated route to contest, with a timeline attached. And an honest line about what the system does not show, because every transparency tool has a boundary, and the boundary is where the interesting material lives.
What has been announced is real and partial at once: a window into the engine, opened first for customers in good standing. Keep the window. Then explain the turnstile, because as it stands the right to know you have been silenced is being handed out by posting volume, and that is not how rights work.
Raffaele Di Marzio
All my “insane” books on cybersecurity and governance are here 👇 https://cyberium.limited/bookshelf.html