X folded its 2020 manipulated media rule into one policy

The persona

This entry is not about one character. It is about the platform rule that governs every account posting a manipulated or synthetic image or video on X, including an account a studio operates on behalf of a fictional persona such as a CGI model or virtual singer. X (formerly Twitter) is the operating company; a persona's account on X is, in every case this policy addresses, run by whoever built and posts as that character, not by the character itself.

What the documents establish

X introduced a dedicated rule for this content on 4 February 2020, described in the company's own announcement. The post states that Twitter (as it was then named) ran a public survey drawing more than 6,500 responses and says more than 70 percent of respondents found 'taking no action' on misleading altered media unacceptable, and that close to 9 in 10 called warning labels an acceptable response; these are the company's own reported survey figures, not an independently audited count. As retrieved on 16 September 2026, X's current policy text folds that 2020 rule into a broader 'Authenticity' page. It now separates 'Manipulated' media, defined as content 'significantly and deceptively altered' in a way that changes its meaning, from 'Out-of-context' media, shared with false claims about its source or subject, and it lists enforcement options ranging from labels to removal depending on severity and history.

Character versus company

A virtual persona's account can only violate this rule through what its operator posts, edits, or fails to disclose; the fictional character has no independent capacity to deceive. The 2020 announcement frames the rule as something 'we' (Twitter, the company) built after consultation, and the 2026 text still describes what 'we' will do, not what any character says about itself. Where a persona's account shares altered media as part of its normal storytelling, this policy asks whether that content is presented as authentic in a way that could cause confusion or harm, a judgment the company makes about the operator's posting choices, not the persona's fictional biography.

What to watch

The current page does not name virtual influencers or CGI personas as a category, so nothing here confirms how X applies the rule to a studio-run account specifically; that gap is worth tracking as an editorial matter, not a documented finding. The 2020 figures also describe attitudes toward the policy's design, not measured enforcement outcomes since. Readers should watch whether X publishes updated survey data or transparency figures tied to this specific rule.

• Does the operator of a given persona account disclose when content has been substantially altered?
• Has X applied a label, and on what basis, when a persona's post could be read as depicting a real event?
• Is there any published enforcement data specific to studio-run accounts, as opposed to individual users?

Taken together, the two documents show a rule that began with a public consultation in 2020 and now lives inside a wider authenticity framework in 2026, but the text itself still leaves the harder question of how it applies to any specific fictional persona's account to the company's own case-by-case judgment.