The voice can change.
Ask for evidence.
AI can make a scam’s presentation more convincing. Familiar voices, fluent writing and polished imagery deserve the same independent checks as any other claim.
Familiar voices, unfamiliar requests
The FTC describes family-emergency scams using cloned voices. A recognizable voice is not enough to authenticate an urgent request. Reconnect with the person through a number you already trust.
Manufactured authority
The FBI describes criminals using generated images, documents and audio to support impersonation. Look for corroboration outside the message or website making the claim.
More convincing messages, at greater scale
The FBI also documents AI-assisted text, translations and fraudulent profiles. Good grammar and personal detail no longer make useful guarantees of legitimacy.
Our focus
Document the deception. Trace the evidence. Keep the lesson useful as the tools change.
Be precise about AI.
The cases currently in this collection do not establish AI generation. Their lessons about impersonation, payment requests and unverified authority remain relevant. A contribution about an AI-enabled incident should link to evidence of the AI involvement, not infer it from a polished design.
Share a documented AI-related case →
Source guidance checked 14 September 2026. This page describes documented patterns, not a prediction of every future attack.