What we found
The Federal Trade Commission has published a National Preparedness Month alert saying that after emergencies, scammers often pose as government officials, charities, or contractors. The alert ties this to recent violent storms, tornadoes, wildfires, heat waves and extreme flooding, and states that scammers are looking to steal people's money and personal information. Among the steps it lists are checking what your insurance covers, keeping scanned documents behind a strong password and multifactor authentication, doing a household inventory, and checking out contractors before you need one. Vet any contractor before disaster strikes: ask people you trust for recommendations, then search the company name online together with the words "scam" or "complaint" — and do it today, not while standing in front of a damaged roof. The advisory does not describe any specific incident, name any company or impersonated agency, say how first contact is made, or give any figure for how many people were targeted or how much money was lost. Our harvested material contains a single official preparedness advisory and no reported cases, so nothing here shows how an approach actually plays out in practice. LIVEFRAUD ALERTS is independent and not affiliated with any government agency, and this alert was assembled only from the harvested source listed above.
- Single source, but it is an official consumer protection agency advisory published under its own name.
- The advisory is general and forward-looking preparedness guidance, not a report of an observed active campaign.
- No incident counts, loss figures, contact channels or impersonated entities are given, so the shape of any live approach is unknown.
- Reviewed by 4 models, 3 from independent houses.
THE RULE
After a storm, fire or flood, someone at your door or on your line may claim to be an official, a charity or a repair contractor — and the FTC says that is exactly when impostors show up.
What we don’t know
- How impostors make first contact — door, phone, text, email or social post — is not stated.
- What payment method or account details are requested is not stated.
- Whether any specific agency, charity or contractor brand is being impersonated is not stated.
- How many people have been approached, and any dollar losses, are not stated.
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 · MODERATE
The single evidence row supports only the core sentence (FTC alert; impostors posing as government officials, charities or contractors). Everything else in the finding — the tie to 'recent violent storms, tornadoes, wildfires, heat waves and extreme flooding', the statement that scammers seek money and personal information, and the enumerated steps (insurance check, scanned documents behind a password plus multifactor authentication, household inventory, vetting contractors) — has no corresponding evidence row. On the material supplied, those specifics are unsourced attributions to the advisory.
DISSENT, PRESERVEDI would publish the first sentence and the caveats and cut or separately source the rest. As written the alert asserts specific advisory contents that the harvested row does not evidence, and that is the kind of small drift that erodes trust in the whole series.
DISSENT, PRESERVEDOn grade: I hold moderate because this is a preparedness advisory with no live campaign, no contact channel and no ask — a reader meeting only this text is not at immediate risk. I want it on record, though, that the underlying real-world pattern (bogus post-disaster contractors taking deposits, and people impersonating relief agencies to harvest identity data from displaced households) routinely produces four- and five-figure losses against victims under duress. Anyone reading my moderate as 'minor' is misreading it.
GOOGLEgemini-flash-latest · HIGH
The finding includes substantial specific details—such as mentions of violent storms, tornadoes, wildfires, heat waves, extreme flooding, insurance checks, multifactor authentication, and contractor vetting advice—that do not appear anywhere in Evidence Row 1.
GROQopenai/gpt-oss-120b · MODERATE
The finding adds details (specific disaster tie‑ins, step‑by‑step advice, contractor vetting guidance) that are not present in the single evidence row, which only states scammers often pose as officials, charities, or contractors.
OPENROUTERopenrouter/free · MODERATE
The finding includes details not in the source evidence—such as linking the alert to specific recent weather events (storms, tornadoes, wildfires, heat waves, flooding), listing concrete protective steps (insurance checks, password protection, household inventory, contractor vetting), and providing specific advice to search for "scam" or "complaint"—none of which are mentioned in the FTC advisory text that was harvested.
Reviewed by 4 independent models; all judged the finding to go beyond the evidence.
▼ Protocol & challenge record
Objection (high): The risk_line invents a contact channel the source does not support: 'someone at your door or on your line' asserts door-to-door and telephone approaches. The harvested row says only that after emergencies scammers 'often pose as government officials, charities, or contractors' — no channel at all. Worse, the draft's own unknowns section states 'How impostors make first contact — door, phone, text, email or social post — is not stated.' The alert therefore contradicts itself: the most prominent, most-shared line asserts exactly what the body says is unknown. Rewrite to channel-neutral language (e.g. 'someone claiming to be an official, a charity or a repair contractor may get in touch').
Not resolved — preserved on the record.
Objection (medium): Source-to-claim stretch in finding sentence 2: 'The alert ties this to recent violent storms, tornadoes, wildfires, heat waves and extreme flooding.' The source uses those events only as a preparedness hook ('a reminder that severe weather and natural disasters can happen anytime'); the impersonation sentence sits in a separate paragraph and is generic ('After emergencies'). The FTC does not report impostors following those particular events. As written, a reader can reasonably infer FTC observed impersonation after specific recent disasters. Reword to 'The alert opens by citing recent violent storms... as a reason to prepare, and separately states that scammers are looking to steal people's money and personal information.'
Not resolved — preserved on the record.
Objection (medium): Audience/targeting mismatch. The callout 'ATTENTION: ANYONE IN A DISASTER-HIT AREA' points the alert at people already hit, but the source is forward-looking preparedness guidance addressed to everyone before an event ('get started now', 'Check out contractors before you need one'). This is the same evidentiary defect the pipeline correctly flagged when it dropped the 'rebuilding' and 'fire' directives under §11 Rule 2 — no harvested source describes disaster-hit readers as a targeted group. The retained callout is inconsistent with those drops. 'ATTENTION: EVERYONE' or a household-preparedness callout is the defensible choice.
Not resolved — preserved on the record.
Objection (medium): watch_icons include 'bank' and 'card'. Nothing in the harvested row mentions payments, bank accounts, cards, wire transfers or gift cards; the draft's own unknowns say 'What payment method or account details are requested is not stated.' Financial-instrument icons assert a payment mechanic that is not in evidence. 'person' is supportable (impersonation); the other two should be dropped or replaced with something evidenced (document/insurance).
Not resolved — preserved on the record.
Objection (low): Attribution gap: the advice sentence about vetting contractors is carried with row_ids: [] and role 'advice', but it is a near-verbatim restatement of the source's own step ('Ask people you know and trust for recommendations. Then search online for the company's name with words like "scam" or "complaint"'). Presenting sourced FTC guidance as unattributed house advice understates the evidence and, given the disclaimer that LIVEFRAUD ALERTS is not affiliated with any government agency, mildly misattributes it. Cite the row.
Resolved: Resolvable mechanically by adding row 222cefa3 to the advice sentence's row_ids; the underlying content check passes against the source text verbatim.
Objection (low): The added flourish 'not while standing in front of a damaged roof' smuggles in the storm-chaser roofing-contractor scenario — a real phenomenon, but one this advisory never describes. It compounds OBJ-1 by painting a specific post-disaster doorstep scene the source does not contain.
Not resolved — preserved on the record.
Objection (low): Date currency / evergreen framing. The source is a National Preparedness Month item dated 2026-09-09; its 'recent' storms and its September framing are time-bound. The claim is written in a bare present tense ('The FTC says...') that will read as a live warning indefinitely. The publication date and the September/National Preparedness Month context should be visible in the claim or lede, not only in the row metadata, so readers can judge currency.
Not resolved — preserved on the record.
Objection (low): Generator hygiene: callout_options and directive_options contain incoherent, token-scraped entries — 'ATTENTION: SCAMMEDS', 'ATTENTION: MULTIFACTORS', 'Send this to any multifactor you know', 'Forward this to the multifactors in your life'. None were selected here, so no live harm, but a slate in which half the options are nonsense provides no real editorial choice and risks selection of a garbage string on a future run.
Not resolved — preserved on the record.
Objection (low): The claim drops the source's hedge 'often' ('they often pose as...') and narrows 'After emergencies' to 'after weather emergencies and disasters'. The finding restores 'often', so this is minor, but the headline claim as worded reads as a flat assertion of current behaviour rather than a generalisation.
Resolved: Partially self-resolved: the first finding sentence retains 'often' and the weather narrowing is consistent with the source's framing. Recommend restoring 'often' to the claim line as a one-word fix; no escalation.
Preserved dissent
ON THE RECORDMy core objection is OBJ-1 and I do not consider it a stylistic quibble. The alert's own unknowns list says the contact channel is not stated, and the risk_line then tells readers to expect someone 'at your door or on your line'. That is not colour; it is an assertion of fact about scammer method, invented at the point of maximum reader attention, contradicted three paragraphs later by the outlet's own caveat. If this ships unchanged I want it recorded that the disclosure block was accurate and the headline risk line was not.
ON THE RECORDI also dissent from the audience_callout. The pipeline correctly refused to tell readers to forward this to people 'rebuilding' or hit by 'fire' because no source describes those people as targeted — and then addressed the whole alert to 'ANYONE IN A DISASTER-HIT AREA' anyway. Either post-disaster readers are an evidenced target group or they are not. Applying the rule to the share directive and not to the callout is not a rule, it is a formality.
ON THE RECORDBroader dissent, for the record: this item contains no incident, no impersonated entity, no channel, no payment method, no victim count and no loss figure. It is a government agency's annual seasonal reminder. Dressing it in alert furniture — a threat callout, a doorstep risk line, bank and card icons — manufactures urgency that the underlying document does not carry. The honest version of this is a short 'FTC publishes annual preparedness guidance' note. I would run it that way or not at all.
The sources
Official sourceThis National Preparedness Month, plan ahead to avoid scams2026-09-09
The Federal Trade Commission has published a National Preparedness Month alert saying that after emergencies, scammers often pose as government officials, charities, or contractors.
Authority: official. Retrieved 2026-09-09.
Limitation: Official advisory only; it names no specific impersonated agency or company, no contact method, and no victim or loss numbers.
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 3 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 2026-09-10.
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