What we found
An FTC consumer alert published 10 September 2026 states that platforms covered by the Take It Down Act are legally required to give people a way to report intimate images shared without their consent, including digitally altered or created images. The same alert says covered platforms must take down the images, and copies of those images, within 48 hours of the request. It directs people to report the content from the post itself — often through the three dots in the upper right corner — or through the platform's help and support pages, with contact details listed in a directory from StopNCII.org. The alert describes StopNCII.org as a free global tool operated by the Revenge Porn Helpline that creates a "digital fingerprint" of an image and shares the fingerprint, not the image, with participating platforms so they can scan for matches and block it from being posted. If a platform does not remove the image within 48 hours, the alert says to report it to the FTC at TakeItDown.ftc.gov. If an intimate image of you is posted without your consent, report it to the platform first, note the date and time you reported it, and if it is still up 48 hours later, take that record to TakeItDown.ftc.gov. Our harvested material does not say how many reports platforms have received, how often images are actually removed inside the 48-hour window, or what enforcement has followed. The source does not name which platforms are covered by the Act or which platforms participate in the StopNCII.org fingerprint scheme. The source also does not describe any specific group of people as being targeted more than others, so we do not claim one. LIVEFRAUD ALERTS is an independent consumer fraud awareness publication and is not affiliated with the FTC, StopNCII.org, the Revenge Porn Helpline or any platform mentioned. This alert was written only from the single official advisory listed in our evidence.
- The single source is an official FTC consumer alert, and every claim here restates its own wording.
- The removal window, the reporting routes and the FTC reporting page are stated explicitly in the source text rather than inferred.
- Only one source was harvested, so scale, enforcement outcomes and platform coverage are not corroborated anywhere in our material.
- Reviewed by 4 models, 3 from independent houses.
THE RULE
Someone whose intimate or digitally altered image is circulating may not know a reporting route and a 48-hour removal duty exist, and may give up after a platform ignores the first request.
What we don’t know
- How many people have used the reporting routes or the FTC page.
- Which platforms fall under the Act's coverage and which participate in the fingerprint scheme.
- What happens, and how quickly, after a report is made to the FTC.
- Whether any particular group is more often affected — the source describes none.
The bench — who voted
4 INDEPENDENT AI MODELS REVIEWED THIS. 3 OF 4 — ONE MODEL DISSENTED. READ WHY.
The card names a count. Here are the seats behind it, with what each one said.
ANTHROPICclaude-opus-5 · HIGH
The finding attributes to the FTC alert a specific reporting procedure — 'report the content from the post itself — often through the three dots in the upper right corner — or through the platform's help and support pages, with contact details listed in a directory from StopNCII.org' — that appears in none of the four evidence rows. Rows 1–4 cover the reporting obligation, the 48-hour removal duty, the StopNCII fingerprint mechanism and the FTC escalation page only. On the stated rule that nothing beyond the listed evidence may be used, this passage is unsourced and should be cut or backed by an explicit row.
DISSENT, PRESERVEDI grade this high on reader risk despite the alert being protective in intent. The underlying harm — non-consensual intimate images, including synthetic ones — is severe, fast-moving and effectively irreversible once distributed, and a reader meets this material at a moment of acute distress. A person who reads '48 hours' as a guarantee, without knowing whether the platform in question is even covered by the Act, may wait two days instead of pursuing parallel remedies. That combination of severe harm plus an unqualified-sounding deadline is what drives the grade, not any defect in the FTC's position.
DISSENT, PRESERVEDBecause the piece names a real removal tool and a real federal URL, it is worth recording that readers searching for these routes are a known target for imitation services charging fees for takedowns. The evidence here says nothing about that and I make no claim on it, but I would not want the panel to treat the absence of such a caution as a sign the topic is low-risk.
DISSENT, PRESERVEDI would not have downgraded the confidence rating on this one. The four sourced rows genuinely carry the central claim; my 'overstated' verdict rests narrowly on the imported procedural detail, and removing that paragraph would move this to 'holds'.
GOOGLEgemini-flash-latest · HIGH
The finding asserts that the alert directs users to report content directly from the post 'often through the three dots in the upper right corner' or via support pages with contact details in a directory from StopNCII.org. None of these specific instructions or directory details appear in the provided evidence rows.
GROQopenai/gpt-oss-120b · MODERATE
Finding adds UI instructions ('three dots in upper right corner') and mentions a directory of contact details from StopNCII.org, which are not present in the cited evidence rows.
OPENROUTERopenrouter/free · HIGH
The finding adds reporting mechanics not present in the supplied evidence rows: reporting from the post through three dots, using platform help pages, and relying on a StopNCII.org directory. Those details are not established by rows 1–4.
DISSENT, PRESERVEDThe central claim that covered platforms must provide a reporting route, remove qualifying images and copies within 48 hours, and that FTC reporting is available if they do not is supported by rows 1, 2, and 4. The overstatement concerns additional procedural details, not the core allegation.
Reviewed by 4 independent models; all judged the finding to go beyond the evidence.
▼ Protocol & challenge record
Objection (high): Publication-blocking defect in the generated furniture: two of the four directive_options are broken placeholder text — "Send this to any discover you know." and "Forward this to the discovers in your life." This looks like a templating bug that scraped the word "discover" out of the source line "If you discover that someone shared intimate images of you...". Any of this reaching a reader destroys credibility on an alert whose whole value is that the reader trusts the procedural instructions. These two options must be deleted, and the pipeline that produced them should be treated as suspect for the rest of the fields it generated.
Resolved: Delete both malformed directive_options before any publication path, and have the Desk spot-check the remaining generated fields (callout_options, watch_icons) from the same generator for similar corruption.
Objection (medium): Action-URL risk is unmanaged. The alert tells readers to go to "TakeItDown.ftc.gov" — a bare hostname the checker has not independently confirmed resolves to an FTC-controlled page. Two compounding hazards: (a) lookalike/typosquat domains are the standard predation pattern around exactly this kind of victim traffic, and this is a fraud-awareness publication that should know better than to publish a naked hostname as an action step; (b) there is a separate, well-known, similarly named service — NCMEC's "Take It Down" (takeitdown.ncmec.org) for under-18 imagery — which readers will confuse with the FTC page. The alert should instruct readers to reach the reporting page by navigating from consumer.ftc.gov rather than typing a hostname, and should not be published until the URL is confirmed live and FTC-controlled.
Not resolved — preserved on the record.
Objection (medium): The 48-hour duty is presented as unconditional, which it is not in the underlying statute. The Take It Down Act's removal clock runs from a *valid* request that meets statutory content requirements (identification of the depicted person, identification/location of the image, a good-faith statement of non-consent, and contact information/signature). The FTC alert compresses all of that to "your request", and the finding inherits the compression — so does the claim line ("must remove them within 48 hours"). A reader who files a bare report and then presents to the FTC on hour 49 may have no valid clock at all. Since the harvested row genuinely does not describe request requirements, the honest fix is a limitation sentence — "the source does not describe what a removal request must contain for the 48-hour clock to start" — not silence.
Resolved: Add a limitation sentence to the finding: "The source does not describe what a removal request must contain, or what starts the 48-hour clock." Soften the claim line from "must remove them within 48 hours" to "must remove them within 48 hours of a request" to track the source's own wording.
Objection (medium): Age-routing gap with real harm potential. The alert directs all readers to StopNCII.org, which in practice serves adults; imagery of minors is routed elsewhere. The audience_callout options include "ATTENTION: PARENTS" and the source's own topic tags include "Protecting Kids Online", so a parent of a minor is a foreseeable reader of this alert. The harvested row says nothing about age eligibility, so the finding cannot assert a minors route — but it also should not imply StopNCII is the universal route. Add to unknowns/limitations: the source does not say whether the StopNCII route applies to images of minors.
Resolved: Add to unknowns: "Whether the StopNCII.org route applies to images of people under 18 — the source does not say." Drop "ATTENTION: PARENTS" from callout_options unless a minors route is sourced.
Objection (medium): Date currency is asserted, not verified. The row is dated 2026-09-10 on a /consumer-alerts/2026/09/ URL. The Desk should confirm this publication date is actually in the past relative to the checking date; a forward-dated official row is the signature of either a harvest error or a fabricated row, and the entire finding rests on this single row with confidence set to "high". If the date cannot be confirmed as past, confidence should drop and the item should hold.
Not resolved — preserved on the record.
Objection (low): Minor source-to-claim stretch: the source says "You might find it by tapping the three dots in the upper right corner"; the finding upgrades this to "often through the three dots in the upper right corner". "Might" to "often" is a small frequency claim the source does not make. Use the source's hedge.
Resolved: Restore the source hedge: "often" to "you might find it".
Objection (low): The advice sentence adds an instruction not in the source — "note the date and time you reported it". It is sensible and low-risk, and it is correctly tagged role: advice with empty row_ids, but it should be visibly framed as the publication's own practical suggestion rather than reading as continued FTC guidance, since the surrounding sentences are all FTC restatement.
Resolved: Attribute the record-keeping step to the publication rather than letting it read as FTC guidance.
Objection (low): Under-harvesting, self-acknowledged but understated. The source page carries a Related Item, "What will the FTC's enforcement of the TAKE IT DOWN Act mean for you?", which is a visible, same-publisher, directly on-topic document covering exactly the enforcement and coverage questions the finding lists as unknowns. Saying "our harvested material does not say" is true but slightly self-serving when the gap was one link away on the harvested page. Either harvest it or note that an adjacent FTC item on enforcement exists and was not harvested.
Not resolved — preserved on the record.
Objection (low): audience_callout "ATTENTION: ANYONE WHOSE PHOTOS WERE SHARED" is broader than the subject matter (it reads as any shared photo, not nonconsensual intimate imagery) and simultaneously narrower ("photos" drops video and drops synthetic images that were never photos of the person). "ATTENTION: TARGETS OF FAKE INTIMATE IMAGES" from the options list is closer, though it drops real imagery. Neither option cleanly covers the source's scope: real and synthetic, photo and video.
Not resolved — preserved on the record.
Objection (low): "high" confidence is defensible for pure restatement of a single official source, but the third confidence_reason ("Only one source was harvested, so scale, enforcement outcomes and platform coverage are not corroborated") is a reason *against* confidence sitting inside the list of reasons *for* it. That reason belongs in limitations. The rating should be explicitly scoped: high confidence that the FTC said this, no confidence about how it operates in practice.
Resolved: Move the single-source caveat out of confidence_reasons into the limitations block and state the scope of the high rating explicitly.
Preserved dissent
ON THE RECORDI would not clear this for publication in its current state. Not because the substance is wrong — the restatement of the FTC alert is faithful, sentence for sentence, and the limitations are honestly drawn. I would hold it for two reasons that have nothing to do with the sourcing quality. First, the directive_options contain broken placeholder output ("Send this to any discover you know"), which means the generation stage malfunctioned on this item; I do not trust the other generated fields until someone has looked at them, and neither should the Desk. Second, we are telling a person in acute distress to type a bare hostname into a browser, on a page published by a fraud-awareness outlet, in a subject area that is a known magnet for lookalike-domain predation, and with a separately branded NCMEC service of almost the same name in circulation. That is the one place where being slightly wrong causes the specific harm we exist to prevent. Confirm the URL or route readers through consumer.ftc.gov.
ON THE RECORDOn the substance I have one standing disagreement with the 'high' confidence framing. High confidence that the FTC published these words is correct. But the claim line, as written, tells a reader that platforms 'must remove them within 48 hours' full stop, and the statute does not work that way — the clock is tied to a request that meets specific requirements the alert never states. Our reader is being handed a countdown they may not actually have started. Faithfully reproducing a source's simplification is still publishing a simplification, and here it has a concrete failure mode: someone waits 48 hours, escalates, and finds the clock never began. A one-line limitation costs us nothing and covers it.
The sources
Official sourceWas your intimate image shared online without your consent? Here’s what to do2026-09-10
An FTC consumer alert published 10 September 2026 states that platforms covered by the Take It Down Act are legally required to give people a way to report intimate images shared without their consent, including digitally altered or created images.
Authority: official. Retrieved 2026-09-11.
Limitation: The advisory does not list which platforms are covered or how the requirement is applied in practice.
Open the original source →Official sourceWas your intimate image shared online without your consent? Here’s what to do2026-09-10
The same alert says covered platforms must take down the images, and copies of those images, within 48 hours of the request.
Authority: official. Retrieved 2026-09-11.
Limitation: No data in the source on how often the 48-hour window is met.
Open the original source →Official sourceWas your intimate image shared online without your consent? Here’s what to do2026-09-10
The alert describes StopNCII.org as a free global tool operated by the Revenge Porn Helpline that creates a "digital fingerprint" of an image and shares the fingerprint, not the image, with participating platforms so they can scan for matches and block it from being posted.
Authority: official. Retrieved 2026-09-11.
Limitation: The source does not name the participating platforms or say how effective the matching is.
Open the original source →Official sourceWas your intimate image shared online without your consent? Here’s what to do2026-09-10
If a platform does not remove the image within 48 hours, the alert says to report it to the FTC at TakeItDown.ftc.gov.
Authority: official. Retrieved 2026-09-11.
Limitation: The source does not describe what happens after a report is filed with the FTC.
Open the original source →Other checks
Published under standing founder pass (A9) — every claim source-mapped by the machine.
▼ What the machine checked
- ✓ Not a community submission.
- ✓ No entity is named.
- ✓ All 5 material sentence(s) map to FTC.
- ✗ anthropic returned "overstated"; google returned "overstated"; groq returned "overstated"; openrouter returned "overstated" — published on the receipt, not blocking (A9 amendment).
- ✓ No audience band is set.
No human affirmed these. They were verified by the classifier described in Amendment A9, on 2026-09-11.
Something wrong here? Tell us and we'll correct it — corrections are published, not quietly edited.
Phishy? Send it → sharelivefraud.com/squire-it
Not affiliated with any government agency, credit bureau, bank, platform, or law-enforcement agency. Informational only — not legal or financial advice.
Naming a source is not an endorsement, and being named here is not an accusation against any company.
Powered by SquireIt™
Verify this receipt at squireit.com