What we found
An FTC consumer alert dated 28 August 2026 states that many people experience fraud but never report it, and that scammers count on you staying silent. The alert says a report filed with the FTC goes into the Consumer Sentinel Network, described as a secure national database available to local, state and federal law enforcement. According to the same alert, FTC investigators, lawyers and analysts review those reports to spot patterns and build cases, while other staff use them to find new scams to cover in Consumer Alerts. The alert directs readers to report what happened at ReportFraud.ftc.gov and to subscribe to the FTC's Consumer Alerts, and links a video titled Why Report Fraud. This harvested alert does not describe any single scam: it names no contact method, no impersonated organisation, no payment channel and no dollar amounts. It also gives no figures for how many scams go unreported, and does not say what happens to an individual report after it is filed. Do one thing this week: describe out loud to one person in your life exactly what the scam you saw looked like — the opening line, the ask, the payment method — so they recognise it if it reaches them next. LIVEFRAUD ALERTS is independent and not affiliated with the FTC or any agency named here; this check pass was built only from the single published consumer alert cited above.
- The underlying material is a single official FTC consumer alert with a publication date, so what the agency said is well supported.
- Only one row was harvested — there is no second source and no independent reporting to corroborate or extend it.
- The row is general reporting guidance rather than a description of a specific scam, so no mechanic, target group or money flow can be stated from it.
- Reviewed by 3 models, 2 from independent houses.
THE RULE
The gap here is silence: the FTC alert says scammers rely on people not reporting, which leaves the same approach free to run through the same circles unrecognised.
What we don’t know
- How many scams go unreported, and over what period — the alert gives no figure.
- Which scams or approaches this alert was prompted by, if any.
- What an individual reporter can expect after filing, including whether money is ever recovered.
- Who is most affected — the alert names no age group, location, income band or family role.
The bench — who voted
3 INDEPENDENT AI MODELS REVIEWED THIS. ALL 3: LOW RISK.
The card names a count. Here are the seats behind it, with what each one said.
ANTHROPICclaude-opus-5 · LOW
Several specific statements are attributed to the alert that the single evidence row does not contain: that 'scammers count on you staying silent', that the Sentinel database is 'secure' and available to 'local, state and federal' law enforcement, that 'FTC investigators, lawyers and analysts review those reports to spot patterns and build cases', that other staff mine reports for future Consumer Alerts, the invitation to subscribe to Consumer Alerts, and the existence of a linked video titled 'Why Report Fraud'. Row [1] supports only the unreported-fraud statement, entry into the Consumer Sentinel Network available to law enforcement, telling someone you know, and the ReportFraud.ftc.gov referral. Everything beyond that is unverified against the supplied evidence.
DISSENT, PRESERVEDThe underlying advice — tell someone and file at ReportFraud.ftc.gov — is accurate, low-harm and points to the correct official channel, so the reader-facing danger here is low even though the sourcing is padded. My 'overstated' verdict is about attribution discipline, not about the advice being wrong.
DISSENT, PRESERVEDIf the harvested alert text genuinely contains the staff-review description and the 'Why Report Fraud' video, the fix is to add those lines to the evidence rows rather than to soften the finding. As presented, a reviewer restricted to row [1] cannot confirm them.
GROQopenai/gpt-oss-120b · LOW
The finding adds that "scammers count on you staying silent," which the source does not state.
OPENROUTERopenrouter/free · LOW
Finding adds claims not present in the source: that scammers 'count on you staying silent', and that FTC investigators, lawyers and analysts review reports to spot patterns and build cases, while other staff use them to find new scams to cover in Consumer Alerts. These details are not in the provided evidence.
Reviewed by 3 independent models; all judged the finding to go beyond the evidence.
▼ Protocol & challenge record
Objection (high): Internal contradiction between finding and limitation. The limitation sentence says the alert 'does not say what happens to an individual report after it is filed', but the finding's own sentences 2 and 3 state exactly what the alert says happens: the report enters the Consumer Sentinel Network, is available to law enforcement, and is reviewed by FTC investigators, lawyers and analysts to spot patterns and build cases. As written the check contradicts itself in the same paragraph and understates the source. If the intended point is that the alert does not say whether the reporter is contacted, whether their specific case is investigated, or whether money is recovered, it must be narrowed to say that.
Not resolved — preserved on the record.
Objection (high): Mis-mined entity in directive_options: 'Send this to any investigator you know' and 'Forward this to the investigators in your life' appear to have been generated from the word 'investigators' in the source, where it refers to FTC staff who review reports — not to any reader group, and certainly not to a group being targeted by fraud. These options invite a directive that has no basis in the row at all, and would misrepresent the alert if selected. Same defect, different route: callout option 'ATTENTION: PARENTS' has zero support in the row (no age group, family role, or household mechanic is mentioned). This is the exact failure the draft itself claims to be guarding against under §11 Rule 2, so the option lists contradict the check's own suppression rationale.
Not resolved — preserved on the record.
Objection (medium): The targeting_dropped entry is malformed and, if surfaced, unintelligible: 'DIRECTIVE NOT IN EVIDENCE: recognise — no harvested source describes this group.' 'recognise' is a verb lifted from the source sentence, not a group. The dropped-directive note therefore documents nothing checkable and looks like a token-level artefact rather than a reasoned suppression.
Not resolved — preserved on the record.
Objection (medium): Advice overreach beyond the source, unattributed. The alert says only 'Tell one person in your life what the scam looked like so they recognize it.' The draft's advice adds 'the opening line, the ask, the payment method' — a three-part script the FTC never wrote — and the advice line carries no row_ids, so a reader cannot tell that these specifics are the publisher's invention. The advice also silently drops the alert's own primary ask (file the report), inverting the emphasis of the source it is built on.
Not resolved — preserved on the record.
Objection (medium): risk_line overclaims. 'Leaves the same approach free to run through the same circles unrecognised' asserts a specific propagation mechanic — a defined 'approach' moving through defined social 'circles' — when the alert names no approach, no channel and no population, as the draft's own limitation concedes. The phrase reads as harvested threat intelligence when it is editorial inference. 'The gap here is silence' plus a direct paraphrase of the FTC's 'scammers count on you staying silent' would be defensible; the propagation clause is not.
Not resolved — preserved on the record.
Objection (low): Source-to-claim stretch on 'other staff'. The row says 'FTC staff also look at reports to find new scams to talk about', which does not establish a second, distinct group of staff. 'While other staff use them' invents an organisational division. Use 'FTC staff also' as the source does.
Resolved: Low materiality and non-misleading as to substance; fix by editing three words ('while FTC staff also look at reports to find new scams to cover in Consumer Alerts').
Objection (low): Video title asserted rather than quoted. The row says 'Watch the FTC's Why Report Fraud video below' and offers a YouTube link; it does not present a rendered title string. 'Links a video titled Why Report Fraud' is a fair reading but is one degree firmer than the evidence, and the harvested text does not include the video URL itself, so the 'link' watch icon is only weakly supported.
Resolved: Resolvable by attribution wording: 'the alert points readers to the FTC's Why Report Fraud video'. Drop the 'link' icon unless the video URL was actually harvested.
Objection (low): Confidence is mislocated. The claim under check is purely 'an FTC alert dated 28 Aug 2026 says X' — a proposition fully settled by a single primary official document on the correct domain (consumer.ftc.gov/consumer-alerts/2026/08/...) whose URL path matches the stated date. Single-sourcing is not a real weakness for a 'what the agency said' claim; it only limits what can be said about the underlying fraud. Grading the whole pass 'moderate' on single-source grounds imports uncertainty about the world into a claim about a document, and may read to users as doubt about whether the FTC published this.
Resolved: Partly self-mitigated: confidence_reasons state plainly that 'what the agency said is well supported' and locate the weakness in the absence of scam specifics, so an attentive reader can see the moderate grade is about scope, not authenticity. Recommend the label stay but the first reason be promoted to lead.
Objection (low): Date currency is asserted, not verified as live. The check records a feed publication timestamp but nothing about whether the page was re-checked, updated, or is still reachable at the time of publishing. For evergreen guidance pages this matters less than for scam-specific alerts, but the pass claims to be 'built only from the single published consumer alert cited above' without an access/retrieval date.
Not resolved — preserved on the record.
Objection (low): Newsworthiness framing. This is evergreen FTC reporting guidance with no incident, no mechanic and no numbers; packaging it with an 'ATTENTION' callout and a fraud 'risk_line' risks implying to a scrolling reader that a new threat has been identified. The draft partly mitigates this in its limitations, but the alert furniture around it pulls the other way.
Resolved: Substantially addressed in-text: the finding explicitly says the alert 'does not describe any single scam: it names no contact method, no impersonated organisation, no payment channel and no dollar amounts', and unknowns record that no affected group is identified. The disclaimer also disavows FTC affiliation.
Preserved dissent
ON THE RECORDI do not accept that this pass is publishable as drafted. The claim itself is sound and the domain, date and quoted content all match the row — that part is clean. But the check states, in the same paragraph, both that the alert explains what happens to reports (Consumer Sentinel, review by investigators, lawyers and analysts) and that it 'does not say what happens to an individual report after it is filed'. That is not a nuance, it is a contradiction, and it undersells the one piece of substance the source actually contains. Narrow it to reporter-level outcomes (contact, case status, money recovery) or delete it.
ON THE RECORDThe option lists are worse than the finding. 'Send this to any investigator you know' and 'Forward this to the investigators in your life' are machine artefacts built from an FTC job title, and 'ATTENTION: PARENTS' is targeting invented from nothing. A pass that suppresses one directive on §11 Rule 2 grounds while offering three unsupported ones has not applied its own rule; it has performed it.
ON THE RECORDOn risk_line: 'the same approach' and 'the same circles' assert a mechanic and a social network that the draft elsewhere admits are absent from evidence. Cut the clause. The honest risk line is that the FTC says scammers rely on silence, full stop.
ON THE RECORDI would also grade authenticity of the document high and scope low rather than flattening both into 'moderate' — as it stands the label invites readers to doubt whether the FTC published this, which is the one thing here that is not in doubt.
The sources
Official sourceDid you spot a scam? Here’s what to do next2026-08-28
An FTC consumer alert dated 28 August 2026 states that many people experience fraud but never report it, that reports enter the Consumer Sentinel Network available to law enforcement, and that readers should tell someone they know and report at ReportFraud.ftc.gov.
Authority: official. Retrieved 2026-09-01.
Limitation: The alert is general guidance about reporting; it does not describe any particular scam, contact method, impersonated brand, payment channel, affected group or loss figure, and gives no statistics behind 'many people'.
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 4 material sentence(s) map to FTC.
- ✗ anthropic 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 2026-09-01.
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