CodeWord
The AGI era · scam judgement desk

A voice is no longer proof.
A code word is.

Cloning a voice now takes seconds of audio, so "it sounded exactly like her" has stopped being evidence. This site does not try to teach you to hear the difference — that advice expires with every model release. It gives you the thing that does not expire: a family code word, a callback on a number you already had, and sixty seconds. Free, nothing uploaded, nothing sold.

If you remember one sentence: anyone asking for money, a verification code, or a transfer to a new account gets hung up on and called back on a number already in your phone. "Don't hang up" and "don't tell anyone" are the two clearest warning signs there are.

Three judgements, each with dated sources

QuestionVerdict in one lineBasis
Can you tell an AI voice clone by ear?No — and trying is the mistake. Verify insteadVerification-first public safety guidance; parent-site capability tracker
Are AI voice robocalls illegal?In the US yes, but not because the voice is syntheticFCC Declaratory Ruling, 8 Feb 2024; FTC Impersonation Rule, 1 Apr 2024; EU AI Act Art. 50 from 2 Aug 2026
I already paid — what now?The first hour decides it. Payment provider first, then reportPayment-rail reality; FBI IC3 2025 Internet Crime Report

Three tools, none of them upload anything

Before it happens

Family code word card

Agree one question only your household can answer, add the two-questions-and-a-callback rule, print a card. The answer is never uploaded and never goes in the link.

Make one →
While it happens

Before you pay

Six questions and a timer. Any one unanswered means do not pay. Printable for the back of a phone.

Open →
When unsure

Scam check, 12 questions

Count the red flags: walk away, be careful, or probably fine. Each flag tied to a named rule. The verdict permalinks so you can send it to family.

Start →

Official routes only

Reporting in the US (IC3, FTC) and the EU/UK, plus the primary sources behind every rule quoted here. This site lists no paid "recovery" services, no security products and no courses — people who have just lost money are the exact target of the second wave. Official help →


62.5 / 100

The AGI-2027 Thesis Tracker — our parent site scores the "AGI by 2027" predictions one at a time, against evidence. It is upstream of this site: each step up in generative capability is another step down for "seeing is believing".

Reading as of 2026-09-06; the parent site is authoritative. agiscorecard.com/progress-index · EU AI Act: what applies now · 中文读者:识骗