What we found
An FTC consumer alert published 7 October 2026 says that possessing or sharing explicit images of children can be a serious crime, and that this includes AI-generated deepfakes. The alert tells parents to report the perpetrator to their local police department, to the National Center for Missing and Exploited Children's CyberTipline on 1-800-843-5678, and to the FBI on 1-800-CALL-FBI or at a local field office. It states that the Take It Down Act requires covered platforms to give parents an easy way to request removal of explicit images of a child, usually through a report option on or near the image, or through the platform's help center if you have no account. If a covered platform does not take the image down within 48 hours, the FTC says to report that platform at TakeItDown.ftc.gov. The alert also points to NCMEC's Take It Down service, which creates a hash — described as a digital fingerprint of the image — to help suppress it from appearing in other places online. If an explicit image of your child appears online, make the report first and the removal request second — call the CyberTipline on 1-800-843-5678 before you spend time arguing with the platform, because the report is what reaches law enforcement. Our harvested material is a single official alert and does not say how the images reach these platforms, who is sending them, or how many families have been affected. The source does not list which platforms count as covered under the Take It Down Act, and we have no harvested material showing what the FTC does after a non-removal is reported. LIVEFRAUD ALERTS is independent and not affiliated with the FTC, the FBI or NCMEC; this alert summarises one published official advisory and nothing more.
- The single harvested row is an official FTC consumer alert with a publication date, and every claim sentence restates its wording directly.
- Phone numbers, the 48-hour window and the reporting address are quoted from the advisory rather than inferred.
- Confidence covers only what the advisory states: there is no second source, no enforcement data, and no description of the people sending these images.
- Reviewed by 3 models from independent houses.
THE RULE
A child's explicit or deepfaked image can spread across platforms within hours, and delay in reporting is what lets copies outrun removal.
What we don’t know
- Which platforms are 'covered' under the Take It Down Act.
- What the FTC does once a platform is reported for missing the 48-hour window.
- How these images are typically obtained or generated, and by whom.
- How many children or families this advisory responds to.
- Whether the hash-based suppression service works across platforms that are not covered.
The bench — who voted
3 INDEPENDENT AI MODELS REVIEWED THIS. 2 OF 3 — ONE MODEL DISSENTED. READ WHY.
The card names a count. Here are the seats behind it, with what each one said.
GOOGLEgemini-flash-latest · HIGH
The evidence row does not mention AI-generated deepfakes, specific phone numbers (1-800-843-5678 or 1-800-CALL-FBI), the publication date (7 October 2026), or the specific operational instruction to file the law enforcement report before requesting platform removal.
GROQopenai/gpt-oss-120b · MODERATE
Finding adds that the FTC alert explicitly includes AI‑generated deepfakes and details a specific UI‑based reporting mechanism, neither of which are present in the cited advisory.
OPENROUTERopenrouter/free · MODERATE
Finding asserts that the FTC alert includes AI‑generated deepfakes, but the evidence row does not mention deepfakes.
Reviewed by 3 independent models; all judged the finding to go beyond the evidence.
▼ Protocol & challenge record
Objection (high): The advice sentence invents a sequencing instruction the source does not give and partly contradicts. The alert says: 'Report the perpetrator immediately to criminal law enforcement AND take action to get online platforms to remove the images as quickly as possible' — two parallel, both-urgent tracks. The draft converts this into 'make the report first and the removal request second ... before you spend time arguing with the platform.' The 48-hour removal clock only starts once a request is made, so telling a parent to defer the platform request is operationally worse than the source's own guidance. It also demotes 'Contact your local police department,' which the alert lists first, in favour of the CyberTipline, and justifies it with an unsourced rationale ('the report is what reaches law enforcement') — local police are law enforcement. row_ids is empty, correctly flagging it as unsourced, but an unsourced sentence that cuts against the source should not ship.
Resolved: Unresolved as drafted. Resolvable by replacing the advice sentence with the source's own parallel framing — e.g. 'The alert says to report the perpetrator immediately and to ask the platform to remove the image as quickly as possible; it does not say to do one before the other' — and dropping 'before you spend time arguing with the platform' entirely.
Objection (medium): The risk_line is an empirical claim with a time quantifier that appears nowhere in the harvested row: 'can spread across platforms within hours, and delay in reporting is what lets copies outrun removal.' The source's only statement on urgency is 'Acting quickly can make a meaningful difference' and 'as quickly as possible.' 'Within hours' and the copies-outrun-removal mechanism are fabricated colour on a single-source item that the draft itself says contains no prevalence or tactics data.
Resolved: Unresolved as drafted. Resolvable by rewriting risk_line to the source's own urgency language ('the FTC says acting quickly can make a meaningful difference') and deleting 'within hours' and the outrun-removal mechanism.
Objection (medium): Two different things both called 'Take It Down' are presented in adjacent sentences without disambiguation: the FTC complaint portal TakeItDown.ftc.gov and NCMEC's separate hash-based Take It Down service (takeitdown.ncmec.org). A parent in distress reading this will plausibly go to the FTC URL expecting hashing, or search 'take it down' and land on a lookalike. The source itself links them separately; the draft flattens them. No URL is given for the NCMEC service at all, while a bare ftc.gov subdomain is given — that asymmetry invites misrouting on a topic where lookalike domains are a known predation vector.
Resolved: Unresolved as drafted. Resolvable by naming the two services distinctly in the same sentence and either supplying both URLs or neither.
Objection (medium): Claim-line stretch: 'should be ... removed by covered platforms, which face an FTC complaint route if they fail to act within 48 hours' implies a standing proactive removal duty. The sourced obligation is to provide an easy request mechanism and to remove within 48 hours of a valid request. The finding body also drops the request-anchor ('does not take the image down within 48 hours' — 48 hours from what?). The source is ambiguous on this point too, but the draft should not inherit and amplify the ambiguity in the headline claim.
Resolved: Partially resolved: the finding does say the Act requires platforms to give parents 'an easy way to request the removal.' Fully resolvable by adding 'of your request' to the 48-hour sentence and rewording the claim line to 'must remove on request.'
Objection (low): The selected audience_callout 'ATTENTION: PARENTS OF CHILDREN ONLINE' is not one of the four callout_options supplied. Either the selection is out of process, or the options list is decorative. Separately, 'PARENTS OF CHILDREN ONLINE' is a near-empty qualifier and narrows the advisory's reach: the alert is addressed to any parent or guardian whose child's image surfaces, which does not require the child to be 'online' at all.
Resolved: Unresolved. Resolvable by selecting 'ATTENTION: PARENTS' from the supplied options.
Objection (low): targeting_dropped contains a malformed string: 'DIRECTIVE NOT IN EVIDENCE: not — no harvested source describes this group.' The word 'not' is sitting where a group name should be. This is a template/placeholder leak and should not reach the public record as-is.
Resolved: Unresolved. Resolvable by fixing or suppressing the string before publication.
Objection (low): Date currency needs a sanity check rather than a transcription. Publication is stamped 2026-10-07 and the Take It Down Act's covered-platform compliance deadline preceded it, so the alert is internally coherent — but the draft asserts the date purely from the harvest metadata with no confirmation that the URL resolves or that the alert has not since been revised. Single-source, single-fetch, no re-check.
Resolved: Partially resolved by the draft's explicit date attribution to the alert itself; residual risk is low and acceptable if the URL is re-fetched at publish time.
Objection (low): Confidence 'high' sits awkwardly beside an advice sentence with no row_ids and a risk_line with no row_ids. The confidence_reasons only defend the claim sentences. If the published artefact includes unsourced operational advice, the confidence label should either be qualified or the unsourced material removed.
Resolved: Resolves automatically if OBJ-1 and OBJ-2 are fixed; otherwise downgrade to 'medium'.
Objection (low): Scope narrowing not flagged as a limitation: the FTC alert frames the Take It Down Act's removal right around parents and children because that is the alert's audience. The statute's notice-and-removal regime is not child-specific. A reader could come away believing the 48-hour mechanism only exists for minors. The draft's limitations cover covered-platform identity and FTC follow-up but not this.
Resolved: Resolvable with one added limitation sentence noting the advisory addresses parents and does not describe the Act's scope beyond that.
Preserved dissent
ON THE RECORDI do not accept 'high' confidence on this item as drafted. The claim sentences are faithful transcription and would merit it on their own, but the published artefact also contains an operational instruction to parents ('make the report first and the removal request second') that has no source, and a risk line asserting spread 'within hours' that has no source. A confidence label attaches to the artefact a reader sees, not to the subset of sentences that happen to carry row_ids. As published, this is medium at best.
ON THE RECORDI regard OBJ-1 as the most serious problem here and I want it on the record even if the Desk waves it through. The FTC alert tells parents to do both things urgently. The draft tells them to do one and defer the other, and supplies its own reasoning for why. In a case where the 48-hour statutory clock does not begin until the platform request is lodged, that reordering can cost a family a day. We should not be improvising triage order for child sexual abuse material reporting off a single consumer alert.
ON THE RECORDI think the two-services-called-Take-It-Down collision (OBJ-3) is a genuine safety defect rather than a tidiness point. Parents under acute distress searching 'take it down' is exactly the search behaviour that lookalike sites exploit. Publishing one URL and not the other makes that worse, not better.
The sources
Official sourceDid someone share an explicit image of your child online? Take these steps2026-10-07
The FTC advisory sets out reporting routes to police, NCMEC's CyberTipline and the FBI, a platform removal right under the Take It Down Act, a 48-hour non-removal complaint route at TakeItDown.ftc.gov, and NCMEC's hash-based Take It Down service.
Authority: official. Retrieved 2026-10-07.
Limitation: One advisory only; it gives procedure, not prevalence, not offender tactics, and not a list of covered platforms.
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.
- ✗ 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 2026-10-08.
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