Quick wins for a faster PC:
Scan for outdated or missing drivers - takes under a minuteDriver Scan →Repair Windows errors before they cause bigger problemsFix Now →“Subtlefakes” is the term used for non-consensual AI images made harder to recognize through relatively slight alterations. They do not have to be wholly synthetic: X’s policy explicitly covers intimate images made by digitally manipulating one person’s face onto another person’s nude body. A 2026 report described a visible trend on X, but the evidence cited here does not establish how common subtlefakes are across the platform or prove that they are “taking over.”
What counts as a subtlefake?
“Subtlefake” is the term Emanuel Maiberg used in an August 20, 2026, 404 Media report about non-consensual AI imagery that can be more difficult to spot after relatively small changes. It is a descriptive term, not a separate category in X’s published rules.
As an Amazon Associate I earn from qualifying purchases.
The underlying harm falls within the broader category of non-consensual intimate imagery, also called non-consensual intimate media or image-based sexual abuse. An image can be abusive even if it combines real and generated material, or if it modifies an existing image rather than generating a person from scratch. X’s policy specifically includes images that superimpose or digitally manipulate someone’s face onto another person’s nude body.
Recommended Free Tools
Consent and context matter. X permits consensually produced adult content when it is marked sensitive; that does not make intimate imagery shared without consent acceptable under its policy.
#1 Best Overall
Does the evidence show subtlefakes are taking over X?
Maiberg’s report describes increasing sophistication and a noticeable trend on X, and quotes image-forensics expert Hany Farid saying large platforms are “really struggling under the weight of this for the first time.” Farid’s comments are an expert assessment, not a prevalence measurement.
The available figures do not establish the share or total number of subtlefake images on X. A 2024 platform-reporting audit tested 50 AI-generated nude images, and an Institute for Strategic Dialogue (ISD) report counted mentions of image-generation tools. Neither is a representative count of subtlefakes or victims across X. “Taking over” should therefore be read as the 404 Media headline’s characterization, not as a quantified platform-wide finding.
Rank #2
How to report a non-consensual intimate image on X
X’s published policy provides an in-app reporting route for unauthorized intimate imagery. Its listed steps are:
Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallCrashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minute- Report the post.
- Choose that it displays a sensitive photo or video.
- Select “Unauthorized photo or video.”
- Select “Unauthorized intimate content.”
- Identify whom the report is on behalf of, then submit it.
X says a report can include up to five posts. Its policy says anyone can report certain listed categories, while some categories may require the affected person or an authorized representative to report. The flow’s request to identify whom the report is on behalf of makes it important to answer accurately.
Rank #3
- Transform audio playing via your speakers and headphones
- Improve sound quality by adjusting it with effects
- Take control over the sound playing through audio hardware
X states that it will immediately and permanently suspend an account it identifies as the original poster of intimate media made or shared without consent. For some other cases, it says it may require removal, temporarily lock the account, and suspend the account after a repeat violation. Those are X’s stated enforcement rules, not a guarantee of a particular response or an independent measure of how consistently the rules are applied.
How reporting routes differ
Platform reporting, a copyright complaint, and a legal removal request are separate mechanisms. The 2024 audit illustrates why the route can matter, but its results apply only to the study’s tested reports at that time.
Rank #4
- Intuitive interface of a conventional FTP client
- Easy and Reliable FTP Site Maintenance.
- FTP Automation and Synchronization
| Route | Who can submit | What it covers | Geography and copies | What the evidence says about outcomes |
|---|---|---|---|---|
| X’s unauthorized-intimate-content report | X allows anyone to report certain listed categories; some categories may require the affected person or an authorized representative. The flow asks whom the report is on behalf of. | X’s policy against intimate imagery made or shared without consent, including digitally manipulated face-and-body composites. | X’s platform policy; the cited policy summary does not specify a jurisdictional limit or promise removal of identical copies through this route. | X describes possible account actions in its policy. The policy is not an independent measure of enforcement results. |
| Copyright report | Not stated in the audit summary. | Copyright claims, not a general substitute for reporting non-consensual imagery. A person should use this route only when they have a valid copyright basis. | The audit summary does not establish jurisdictional scope or whether identical copies were covered. | In a 2024 audit, all 25 images submitted through the copyright-report condition were removed within 25 hours. This result is not a guarantee of current outcomes. |
| U.S. TAKE IT DOWN Act request | The FTC’s May 2026 guidance describes a valid request, but the cited summary does not specify all eligibility or submission requirements. | Covered real intimate imagery and “digital forgeries,” including images created or altered using software, apps, or AI. | U.S. law; covered platforms must remove the reported image and known identical copies after a valid request. | The FTC says removal must occur within 48 hours of a valid request. This is a statutory requirement, not a prediction of how long any particular case will take. |
The same 2024 audit tested 25 reports through X’s non-consensual-nudity route. None of those 25 images was removed during the 21-day study. That experiment is a historical comparison of two reporting channels, not evidence that X will handle every current report the same way.
What the TAKE IT DOWN Act means in the United States
The Federal Trade Commission’s May 2026 guidance says the U.S. TAKE IT DOWN Act covers both real intimate imagery and digital forgeries, including images created or altered with software, apps, or AI. For a valid request, a covered platform must remove the content and known identical copies within 48 hours. The FTC says covered platforms must provide a clear process and may face enforcement action; its guidance lists a potential civil penalty of $53,088 per violation.
Best Value
Those requirements apply in the U.S. and depend on the request meeting the law’s requirements. The FTC guidance is not a description of X’s separate in-app policy route, and the 48-hour deadline should not be treated as a guarantee that every report made through X’s ordinary reporting flow will produce removal within that time.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What is known about the wider tool ecosystem
Individual posts are only one part of the problem. An October 2025 ISD report described tools promoted and sold through adult discussion forums, search engines, 4chan, and X. ISD said 31 active tools for sexually explicit image abuse were readily discoverable on X, 4chan, and common search engines as of May 2025. It also counted almost 290,000 mentions of such tools on X between June 2020 and July 2025; those mentions accounted for 70 percent of mentions across the seven platforms and other sources it examined.
These figures measure tool discoverability and mentions, not how many images were made, how many were shared on X, or how many people were targeted. A September 2026 UC Berkeley Center for Long-Term Cybersecurity white paper examines the broader chain of human creators and distributors, models, deployment, and distribution. Its analysis supports looking beyond individual posters when considering how the abuse is enabled.
Maiberg also suggested that X’s ad-revenue sharing could help explain why some accounts post repeatedly, while noting that promoting monetized accounts elsewhere with subtlefakes was rare in his observation. That is the reporter’s explanation of a possible incentive, not proof of the motives behind every account.
What to do if an image involves you
- Use X’s unauthorized-intimate-content reporting flow and provide accurate information about whom you are reporting on behalf of.
- Keep a record of the post and its account details for your own use. Avoid sharing or reposting the image as a way to draw attention to it.
- If you are in the United States, check the FTC’s TAKE IT DOWN Act guidance for the valid-request process and whether the covered-platform requirements apply to your case.
- If you are in the United Kingdom, Ofcom’s February 3, 2026, update directed people under 18 to Childline’s Report/Remove service and adults to the Revenge Porn Helpline.
Legal routes vary by location. Ofcom said it contacted X on January 5, 2026, and opened a formal investigation on January 12 into whether X had adequately assessed and mitigated the risk of sexual deepfakes spreading and removed them promptly once identified. In its February 3 update, Ofcom said the investigation was ongoing and that X had told it measures had been implemented. That update describes the status at that date, not a final finding.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.




