TikTok Community Guidelines Update, September 2026: What Changed on the 24th, Sentence by Sentence

Tomorrow morning a new version of TikTok's Community Guidelines takes effect, and the claim going around, that the AI and authenticity rules got tougher, does not survive a line-by-line comparison. We put the current text and the September 24 text side by side on September 21, block by block, and the chapter that governs multiple accounts, automation and reposting did not change by a single word.

That is the reassuring half. The other half matters more if you run more than one account: the ban sentence now reaches further than it did, and for the first time the guidelines describe a way back. This is the full TikTok community guidelines update for September, limited to what the text itself says.

The dates, and the four chapters that did not move a word

The version that takes effect tomorrow says it was released on August 25, 2026 and becomes effective on September 24, 2026. The current version was released August 14, 2025 and took effect September 13, 2025. The banner on the live guidelines page still reads, as of September 23: "On September 24, 2026, we are updating our Community Guidelines so you can continue to create, discover, and share safely on TikTok." (tiktok.com/community-guidelines/en/, checked 2026-09-23; the new version is served with the parameter cgversion=2026H2update.)

Both versions list the same chapters. No chapter was added or removed. Here is where the comparison came out identical:

• Safety and Civility: 32 text blocks, all identical.
• Integrity and Authenticity: 23 text blocks, all identical. Two internal element IDs were renamed; the words were not touched.
• For You feed Eligibility Standards: the single block on that page is identical.
• Enforcement: three blocks, all identical.

So the sentences people quote most keep their exact wording after tomorrow. The multiple-accounts sentence still reads "You can have multiple accounts—for example, for fan content or creative expression—but not to deceive others or break the rules." It still sits next to "We strictly prohibit automation tools, scripts, or other tricks designed to bypass our systems." And the Spam list still includes "Using automation to run many accounts or send repetitive content."

Read those three together, not one at a time. Having several accounts is allowed in writing. Using automation to run many accounts is named on its own, because the sentence says or, not and. Nothing in the update makes "my posts are all different" a defence, and nothing in it changes where that line sits.

Accounts and Features: the edits that land on anyone running more than one account

This chapter has eleven blocks, and three edits in it are the reason to read this update at all.

1. The ban can now reach associated accounts. The current sentence: "If someone seriously breaks the rules or tries to dodge enforcement, we may ban all of their accounts." The new sentence adds five words at the end: "...we may ban all of their accounts, including associated accounts." The guidelines do not define "associated". We looked at what that phrase can and cannot mean in our breakdown of the associated-accounts sentence.

2. The reason for a ban went from one behaviour to a category. The old list ended with "or pretending to be someone else". The new one ends with "or inauthentic behavior (e.g. pretending to be someone else)". Impersonation is now an example of the rule, not the whole rule.

3. A written route back, for the first time. The old enforcement line was "That can mean a warning, a restriction, or a full ban, depending on what happened." The new one keeps that range and adds: "In some cases, users may be given the opportunity to complete optional policy training to restore full access to the platform." Note the three hedges: in some cases, may be given, optional. It is not a promise and not a replacement for an appeal. The details are in our piece on the policy training route.

A fourth edit is new too, but it is about protection rather than punishment. When TikTok gets a credible report that a user is at risk of harm off-platform and that a third party may get into their account without permission, "we may suspend or restrict access to the account to help protect the user and our community."

The update did not move the line between allowed and not allowed for multi-account operators. It changed how far the punishment reaches once that line is crossed, and added a way back.

Regulated Goods: the chapter that was actually rewritten

Seven of this chapter's eleven blocks changed. If you post about money, medicine or cosmetic procedures, this is your chapter.

• Three new regulated categories are defined. High-Risk Financial Services ("Products or activities involving the management of money and other financial assets that are predatory, contain extreme risks, or lack transparency"), Invasive Cosmetic Procedures ("Procedures or treatments that involve cutting the skin or using needles to modify one's appearance") and Restricted Medical Services (services that diagnose or treat a condition "where an official license is required to operate").
• They are banned and allowed at the same time, depending on who posts. High-risk financial services and restricted medical or invasive cosmetic services join the Not Allowed list. Invasive cosmetic procedures, restricted medical services and "Financial education and information about financial products and services" join the exceptions that verified businesses can use. Explaining how an index fund works is education; selling a product is marketing, and marketing needs the business credential.
• "Registered business accounts" becomes "Verified business accounts" throughout. The bar for regulated sellers moves from having registered to having been verified. (TikTok's own sign-up line in the new text spells it "verifed" [sic].)
• Tobacco becomes "Tobacco and Vaping Products", and gambling and firearms content moves from a regional age restriction to a general one.
• The consequence of undisclosed commercial content is worded differently. The old text said undisclosed commercial content "will be ineligible for the FYF". The new text: "If we find commercial content that hasn't been properly disclosed, we may reduce its visibility or apply the content disclosure setting." The duty to disclose did not change. Only the wording of what happens if you skip it did. If you are unsure whether a post needs the label, this is where the disclosure line sits.

For brands in these industries the ad side has its own gate, covered in our look at TikTok One for restricted industries. The guidelines chapter only governs organic content.

The smaller edits, and the one that only looks new

Outside those two chapters there are a handful of changes. They are smaller, but at least one matters for short-form editors.

The jump-scare line is the only one with a direct effect on everyday creators. Clips that rely on a sudden loud sound or a flash cut to hold the viewer can now be kept out of the For You feed. They stay on your profile, but they stop being recommended.

The verification sentence is the one to be careful with. You will probably see it quoted as "TikTok adds new account verification checks". It is not new. Moving it between chapters does not change what it means.

A checklist to take into September 24

Nothing in the update needs panic. A few things do need a decision:

1. List every account you consider "yours", including old ones, test accounts and accounts a teammate runs from your devices. If one of them gets a serious ban, that list is roughly what "associated" could cover.
2. Keep the five habits that hold up under both versions: log in to your own accounts in your own local browser, never hand over passwords, make each account's content its own, keep a human reviewing what goes out, and post at a human pace. None of them is a compliance guarantee. They are what the text still supports.
3. Audit your hooks for jump scares if you edit short clips, especially AI-generated ones that lean on sudden sound effects.
4. If you post about finance, clinics or cosmetic treatment, decide which of your posts are education and which are marketing. Only verified businesses get the marketing exception.
5. Turn on the disclosure setting for paid content anyway. The softer wording only changes what may happen to you. The duty to disclose is still there.
6. If you are restricted after tomorrow, look for a policy training prompt, but file your appeal as normal. Training is optional and offered only "in some cases".

If you run several accounts with a tool, NoobClaw included, the safer reading of the new ban sentence is to treat your accounts as one unit: one serious strike can now be written up as a strike on all of them.

FAQ

Did TikTok change its rules on AI content in the September 2026 update?

Not in the Integrity and Authenticity chapter, where AI-generated content and authenticity rules live. All 23 text blocks there are identical in both versions. The AI-adjacent change is elsewhere: the new For You feed exclusion for shock effects such as jump scares, which applies whether or not a clip was made with AI.

Does "associated accounts" mean all accounts on the same phone or IP?

The guidelines do not say. The phrase is not defined in either version, and TikTok has not published how it decides that accounts are associated. Anyone giving you a precise list of signals is guessing. What the text does make clear is the trigger: a serious violation or trying to dodge enforcement, not simply having several accounts.

Will completing policy training unban my account?

The text does not promise that. It says that in some cases users "may be given the opportunity" to complete "optional" training "to restore full access". It is offered at TikTok's discretion, and nothing in the guidelines makes it a substitute for an appeal or a guaranteed outcome.