False Claims on X

What the rules ask, and the question they never reach

X's abusive behavior policy is short and unusually candid, which makes it easier to work with than most. It prohibits targeted harassment, described by X as malicious, unreciprocated targeting of individuals, particularly when shared to humiliate or degrade someone. It prohibits incitement of harassment, meaning behavior that encourages others to target specific people. It prohibits denial of verified mass-casualty events where shared in abusive context, unwanted sexual content, and the use of insults or profanity to target others. Misgendering and deadnaming are actioned where required by local law.

Now the paragraph that decides most business complaints. X states that behavior which looks harmful in isolation may not be actionable in context, that posts calling attention to, condemning, or highlighting others for harmful behaviors are not violations where there is no abusive intent, and that criticism of institutions, practices and ideas remains permitted.

A thread accusing a company of ripping people off is criticism of an institution. On the operator's own text, it sits outside the policy. There is nothing to appeal, because nothing was decided against you — the conduct was never within scope. And there is no separate falsity rule waiting behind it: X publishes no defamation policy, and a claim that a post is untrue is not a category the report flow offers.

These pages carry no visible revision date, so what follows is dated by my reading of them on 12 August 2026. X changes its rules and its URLs more often than most operators, as the next section but one demonstrates. Open the live policy before you file anything on the strength of this page.

Three outcomes, and only one of them is a takedown

X lists its enforcement options, and the distinction between them is where most money in this market is quietly made.

“Reducing visibility through removal from search, timelines, and recommendations.”

— X Help Center, Abusive behavior, read 12 August 2026

In its own Q3 2025 report to the New York Attorney General, X defines visibility reduction as removing the post from search results, in-product recommendations, trends, notifications and home timelines. Note what is absent from that list: the post itself. It remains published at its address. Anyone holding the link sees it. Anyone reading the author's profile sees it. Anyone who already screenshotted it has it regardless, and a search engine that indexed the URL may still return it.

The second option is requiring post removal, where the author must delete the post before regaining normal account function — X's own most common route to content actually disappearing, and note that it is the author who deletes. The third is account suspension, which X reserves for accounts dedicated to harassment and for the most serious sexual-content violations, and which can be permanent on a first offense where the conduct is severe enough. Where one person is running several violating accounts, X may require them to choose one to keep and suspend the rest.

When a vendor reports that a post was “actioned” or that its reach collapsed, ask which of the three happened. The first one is real and it is not removal, and for a reputation problem where the damage travels by link and screenshot, it may change very little.

The report a business is most likely to get actioned

Start with a fact that tells you something about the state of published advice on this subject. The impersonation policy URL that most guides still cite, help.x.com/en/rules-and-policies/x-impersonation-policy, returned 404 when I checked it on 12 August 2026. The live rule is the Misleading and deceptive identities policy, and it says this:

“You may not impersonate other identities of individuals, groups, or organizations to deceive others.”

— X Help Center, Misleading and deceptive identities policy, read 12 August 2026

The policy covers fabricated personas built from fake profile photographs, stolen bios and misleading profile information. Parody, commentary and fan accounts are permitted, but conditionally: X requires such an account to carry a label, an avatar distinct from the real entity's, a term such as “parody” or “fan” at the start of the account name, and similar language in the bio. An account that fails those labeling requirements is reportable even where its content is transparently satirical — the violation is the presentation, not the joke.

This is the report most worth a business's attention, because the reviewer can settle it by looking at the account. No fact about the company is in dispute. Where the brand rather than the identity is being misused, X maintains a separate trademark policy, which is a different form and a different team. And X states that authentic coordination — people genuinely organizing to express a viewpoint — is permitted so long as it breaks no other rule, which is the line between a critical campaign and a manipulation network.

Withheld is not removed

Country withholding is the mechanism most advice on this subject gets wrong, usually by describing it as a takedown obtained through legal channels. It is neither.

X withholds content “in a particular country” on “a valid and properly scoped request from an authorized entity,” submitted by a “law enforcement agent, government official, or other third-party,” and “will attempt to notify affected users…via the email address associated with the affected X account(s) and/or through a notification within the X app.”

— X Help Center, About country withheld content, help.x.com, read 12 August 2026

Two consequences follow, and both of them matter to a US company. First, withholding is geographic. The post is not deleted; it is hidden inside one country and remains live at the same address everywhere else in the world, including to anyone in that country using a VPN. Second, the notice X sends identifies the specific content that was reported or withheld and the origin of the request. The author finds out, and learns something about who complained.

That second point generalizes beyond this mechanism. Most enforcement on X is notified enforcement. Anyone imagining a quiet removal, executed without the person who posted ever knowing a complaint was made, is imagining a process that does not exist here. That is not a reason to avoid reporting. It is a reason to preserve the record first and to decide, before filing, what happens if the response is a louder post rather than a deleted one.

Who reports, and why the executive's own account matters

X states that anyone can report violations, and then qualifies it: X sometimes needs to hear directly from the person being targeted before it will enforce. Both halves are true at once and the second one is the operative one.

In practice, a report filed by a communications manager about conduct aimed at a named executive is a weaker report than the same facts submitted from the executive's own account. Nothing about the content changes. What changes is that the platform can see the target confirming that the targeting is unwanted, which is the thing it says it sometimes needs. If an executive is unwilling to file, that is a legitimate decision with a consequence attached, and the consequence should be understood before a strategy is built on reports.

Reporting on X is in-product for policy violations: the menu on the post or profile, then a category, then the prompts. There is no free-text box in which to explain that a claim is false, and there was never going to be. Legal demands travel a different path entirely, through X's separate legal request submission site rather than the in-app flow, and a demand letter sent to the in-app report queue is a letter nobody reads.

Where the published record runs out

This site quotes operators and dates the quotes, which means it also has to say where the operator has published nothing. On X, that point arrives at appeals.

Account-level enforcement can be appealed in-product through the Help Center. Beyond that, I could not confirm from a dated X page, read on 12 August 2026, whether decisions to reduce a post's visibility are appealable at all, or what timescale X applies to any appeal. So this page does not state one. Anyone who quotes you a turnaround for an X appeal should be asked which page it appears on and when they read it, and if the answer is a vendor blog rather than the operator, that figure is worth exactly nothing.

The same restraint applies to volumes. X publishes enforcement reporting, but no figure appears on this page that I did not open and date myself, because a stale statistic repeated confidently is how most of this market's advice becomes wrong without anyone noticing.

What can be said is narrower and more useful. There is no published service level for reports. There is no published defamation process. There is no route that turns on whether a statement about a company is accurate. And where the record is silent, the honest answer is that it is silent, rather than an estimate wearing the clothes of a fact.

What will not work, and what is left

Named plainly, because each of these has been sold to somebody this month:

• Reporting a critical thread as abuse. X's own policy protects criticism of institutions, practices and ideas. The thread about your company is the thing the exception describes.
• Telling X the post is false and waiting. Notice to a platform does not create an obligation to act — Zeran v. America Online (4th Cir. 1997) put distributor liability inside publisher liability, both foreclosed by section 230. “I told them and they left it up” is a lawful outcome, not a wrong.
• Treating a country withholding as a removal. It is hidden in one country and live everywhere else, at the same address.
• Treating reduced visibility as a takedown. The post is still published, and the screenshot was taken before any of this started.

What is left is real, though smaller than the promise. Identity reports, where the question is verifiable from the account. Conduct reports, where a persistent attacker is running multiple accounts. Preservation, done properly and done first, because a post can be deleted by its author at any moment and X keeps nothing for you. And then the work that never depended on X at all: what a search for the company returns, what the company itself publishes, and a documented record of what was posted, by whom, and when.

Frequently Asked Questions

Published 2026-08-12