FRAUD CHECK — Squire It™
sharelivefraud.com/squire-it
LIVE FRAUD ALERT
LIVEFRAUD Check #102
FTC WARNS

Scammers posing as FTC officials call to say a phone was registered in your name for overseas crimes, then hand you to a fake police officer and demand money to avoid "arrest."

HIGH CONFIDENCEPublished 2026-10-23
S
Q
U
I
R
E
D
E
S
K
·
C
H
E
C
K
E
D
·
B
A
T
T
L
E
D
·
R
E
C
E
I
P
T
E
D
·
S

What we found

The FTC published a consumer alert on 22 October 2026 describing callers who claim to be FTC officials and say a phone was registered under the victim's name and number and is being used to commit scams in another country. According to that alert, the caller reads out a badge ID that is fake, and threatens to deactivate the victim's phone number if they do not cooperate. The alert says the caller then transfers the victim to a fake police officer to file a report, and if the victim hesitates, tells them what to search online to find a police department number to call back — where people involved in the scam answer. The FTC says the callers ask for ID and other personal information for the supposed report, instruct the victim to keep everything secret, and later claim the victim is under arrest and must travel to another country unless they send money for a "security deposit." The agency states that a real FTC employee will never ask you to send money, never connect you to someone who says to send money, never threaten to disconnect your phone, never tell you to keep conversations secret, and never text a photo of an employee ID to confirm who they are. If a caller says your phone number is tied to a crime, hang up and do not call any number they give you or tell you to search for — look up the agency yourself and call back on your own schedule. The harvested advisory does not say how many people have been targeted, how much money has been lost, what payment methods the callers request, or which phone numbers or regions are involved. LIVEFRAUD ALERTS is an independent consumer awareness publication and is not affiliated with the Federal Trade Commission or any law enforcement agency; this alert was written only from the harvested material cited above.
THE RULE
High pressure and fear: a fake badge, a fake police transfer and a threat of arrest abroad, ending in a demand to send money.

What we don’t know

The bench — who voted

3 INDEPENDENT AI MODELS REVIEWED THIS. ALL 3: HIGH RISK.

The card names a count. Here are the seats behind it, with what each one said.

ANTHROPICclaude-opus-5 · HIGH
Only one evidence row is supplied, and it attests only the opening premise: that an FTC alert dated 22 October 2026 describes callers impersonating FTC officials who say a phone registered in the victim's name is being used for scams abroad. Everything after that in the finding — the fake badge ID read aloud, the threat to deactivate the number, the transfer to a fake police officer, the instruction on what to search online so the victim reaches conspirators, the request for ID, the secrecy instruction, the claimed arrest, the demand for a 'security deposit' — is not carried by any row. The row's own limitation note says the alert is described 'in general terms,' which if anything cuts against the level of script detail asserted.
DISSENT, PRESERVEDI would not block the safety advice itself. 'Hang up, do not call a number the caller gives you or tells you to search for, look the agency up independently and call back on your own schedule' is correct, low-cost guidance that harms no one who follows it, and the underlying impersonation pattern is real and well documented in general. My objection is to presenting unsourced script detail and unsourced agency quotations as if they were in the cited material — fix the sourcing, keep the advice.
DISSENT, PRESERVEDOn danger: I grade this high and would defend it. Government-impersonation calls that escalate to a fake law-enforcement handoff and an arrest threat exploit fear and urgency, and the secrecy instruction is specifically designed to defeat the family-member or bank-teller intervention that normally stops a transfer. Losses in this category are typically large and irreversible. The high grade reflects the pattern, not the quality of this draft.
GROQopenai/gpt-oss-120b · HIGH
The evidence only confirms the FTC alert mentions callers claiming FTC affiliation and a phone being used in overseas scams; all additional details about badge IDs, threats to deactivate, fake police transfers, secrecy instructions, arrest threats, travel demands, and security deposit requests are not present in the cited source.
OPENROUTERopenrouter/free · HIGH
The finding includes many specific details (fake badge ID, threat to deactivate the phone, transfer to a fake police officer, instruction to search online for a callback number, requests for ID and personal information, secrecy instruction, claim of arrest and required travel, demand for a “security deposit,” the FTC’s list of what real employees will never do, and the advice to hang up and look up the agency yourself) that are not contained in evidence row [1]. Row [1] only states that the alert describes callers claiming to be FTC officials and alleging a phone registered in the victim’s name is being used for scams abroad; it explicitly notes the alert gives no case numbers, loss totals, incident dates, or payment methods. Therefore the finding stretches the source well past what its text says.

Reviewed by 3 independent models; all judged the finding to go beyond the evidence.

▼ Protocol & challenge record
Objection (medium): Modal stretch in the call-back sentence. The source hedges: "If you hesitate, they MIGHT have you call back the police department directly." The finding states it flatly: "if the victim hesitates, tells them what to search online to find a police department number to call back." A conditional, optional step in the FTC's description has been upgraded to a fixed stage of the script. Insert "may" / "sometimes" to match the source.
Resolved: Reword to "if the victim hesitates, the caller may tell them what to search online to find a police department number to call back" — restores the source's hedge.
Objection (medium): Headline claim overstates the money demand. The source says the scammers claim you are ALREADY under arrest for the scams and that you must travel to another country UNLESS you send money for a "security deposit." The claim line compresses this to "demand money to avoid 'arrest.'" That is a different and more generic extortion framing — the distinguishing, searchable detail (told you must fly abroad; money framed as a "security deposit") is lost. The finding body gets it right, so the claim line is the weakest-worded part of the package.
Resolved: Rewrite the claim line to carry the distinguishing detail, e.g. "...then hand you to a fake police officer, claim you are under arrest and must travel abroad, and demand a 'security deposit.'"
Objection (medium): watch_icons includes "bank" with zero evidentiary basis. The draft's own unknowns explicitly state "What payment method the callers ask for" is unknown, and the limitation line repeats that the advisory gives no payment methods. Displaying a bank icon asserts a payment channel the source never names and contradicts the package's own limitation. Strike it.
Resolved: Remove "bank" from watch_icons. No payment channel is evidenced and the package's own unknowns say so.
Objection (low): watch_icons "text-message" is also a stretch. Texting appears in the source only inside the negative guidance ("a real FTC employee will never text you a photo of their employee ID"). Nothing in the harvested script says these scammers actually send texts. An icon derived from a hypothetical in a never-list is source-to-claim drift, though milder than the bank icon.
Not resolved — preserved on the record.
Objection (medium): Defective callout/directive options. "ATTENTION: SCAMMEDS" is malformed and victim-blaming; "ATTENTION: IMPERSONATORS", "Send this to any impersonator you know" and "Forward this to the impersonators in your life" address the perpetrators rather than readers at risk and are nonsensical as consumer guidance. Even unpublished, they are live options that could be selected. They should be removed from the option set, not left on the menu.
Not resolved — preserved on the record.
Objection (low): The source's own action channel was dropped. The FTC alert closes by directing readers to report at ReportFraud.ftc.gov. The draft's advice sentence tells readers only to hang up and look the agency up themselves. The reporting route is the one piece of actionable, fully sourced guidance in the advisory and omitting it weakens the alert without any editorial reason.
Resolved: Add the FTC's reporting route (ReportFraud.ftc.gov) to the advice sentence, attributed to the same row.
Objection (low): The advice sentence "call back on your own schedule" carries no row_id and is not in the source. "Do not call any number they give you or tell you to search for" is squarely supported; "on your own schedule" is generic add-on phrasing. Acceptable as labelled advice, but it should not be allowed to read as if the FTC said it — it sits in the same sentence as sourced guidance.
Not resolved — preserved on the record.
Objection (low): Context dropped: the source frames this as "the latest twist on FTC impersonators," i.e. a variant of an ongoing impersonation pattern. The finding presents the script as a free-standing scheme. Readers lose the signal that FTC impersonation calls are a recurring category and that the phone-deactivation hook is the new element.
Not resolved — preserved on the record.
Objection (low): targeting_dropped contains a garbled entry: "DIRECTIVE NOT IN EVIDENCE: recognise — no harvested source describes this group." "recognise" is not a group. The log line is unreadable and should either name the actual dropped target or be removed; as written it provides no audit value.
Resolved: Either name the group that was actually dropped or delete the line.
Objection (low): Date currency should be confirmed against the publishing clock. The harvested row is stamped 2026-10-22; the finding repeats that date as fact. If the system date is earlier than that stamp, the row is mis-dated or mis-harvested and the whole alert inherits the error. Worth one mechanical check before release — the rest of the package depends on it.
Not resolved — preserved on the record.
Objection (low): Single-source "high" confidence. It is defensible here because the source is the impersonated agency's own consumer alert on its own verified domain (consumer.ftc.gov), which is as close to primary as this claim can get. But the package should not let "high" imply corroborated prevalence: there is no second source, no incidence data, and no independent verification that the script is occurring as described beyond the FTC's assertion. The finding's limitation line covers most of this; confidence_reasons should state plainly that confidence attaches to what the FTC said, not to scale.
Not resolved — preserved on the record.
Preserved dissent
ON THE RECORDThe "bank" watch icon should not ship. The package states in two separate places that the payment method is unknown, then renders a bank icon anyway. That is the package contradicting itself in the most visually prominent element, and readers will read an icon faster than they read a limitation line.
ON THE RECORDI do not accept the flat phrasing of the call-back step. The FTC wrote "might." We wrote a certainty. Small word, but it is exactly the kind of quiet upgrade that turns an accurate alert into an inaccurate one, and we have no second source to absorb the error.
ON THE RECORDThe claim line's "demand money to avoid 'arrest'" is the generic shape of every government-impersonation scam. The specific, memorable hooks here — the fake badge ID read aloud, the threat to deactivate your number, the instruction to search for the police number yourself, the "security deposit" to avoid being sent to another country — are what would actually let a reader recognise this call in progress. Burying all of them below a generic headline is a real loss of protective value.
ON THE RECORD"ATTENTION: SCAMMEDS" and the three directives aimed at "impersonators" should be struck from the record, not parked as options. One of them is a slur against victims produced by what looks like a generation fault; the others instruct readers to forward a fraud warning to fraudsters. Leaving them in the option list means a future pass can select them.
ON THE RECORDI accept "high" confidence. A consumer alert published by the impersonated agency itself, on its own domain, describing the script it is being impersonated in, is the strongest available evidence for this claim and does not need corroboration to support "the FTC says X." My objection is to the icons and the modal, not to the confidence grade.

The sources

Official sourceThe FTC is not calling to deactivate your phone number2026-10-22
The FTC published a consumer alert on 22 October 2026 describing callers who claim to be FTC officials and say a phone was registered under the victim's name and number and is being used to commit scams in another country.
Authority: official. Retrieved 2026-10-23.
Limitation: The alert describes the script in general terms and gives no case numbers, loss totals, dates of individual incidents or payment methods.
Open the original source →

Other checks

Every check we have published →

Share this receipt
sharelivefraud.com/check/GhU37Yg

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"; 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 2026-10-23.

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

Join Squire’s First Watch

Alerts before the feed. Credit when your summons becomes a receipt. A vote on what we check next. Founding names are permanent.

Get the next one

We publish a receipt for every alert, including the ones we decide not to run.

We will ask you to confirm before anything is sent. Your address is used for this and nothing else, and is never shared.