AI scams: a practical response to convincing impersonation

Why a convincing voice can mislead

The FTC warns that scammers can use voice cloning in family-emergency schemes. Its central advice is to verify the story by contacting the person through a known number, or another trusted family contact if necessary. Recognizing the voice is not sufficient verification. Federal Trade Commission: Scammers Use AI to Enhance Their Family Emergency Schemes

Imagine a caller who sounds like a grandchild and says a payment must be made immediately. End the call and use your saved contact information. If you cannot reach the person, contact someone who can independently check. A demand to keep the call secret is part of the situation to question, not a reason to skip verification.

Patterns worth recognizing

An impersonated business may send a changed bank-account instruction. A fake support representative may ask you to install remote-access software. A message may use information from public posts to appear unusually personal. A realistic advertisement may borrow a familiar face to promote a false opportunity.

These approaches existed before generative AI. AI can make some of the content easier to produce or more convincing. The defense is to verify the request and the destination, rather than trying to judge authenticity from grammar or production quality alone.

Set up a simple household procedure

Keep important contact numbers somewhere you can find without searching a suspicious message. Agree with family members that unusual urgent money requests will always receive an independent callback. A private family phrase can be an extra check, but do not make it the only defense; it could be exposed or guessed.

For business payments, confirm a change of bank details with a known contact through an independent channel. Separate the person requesting a change from the person approving a consequential payment when possible. Practice the procedure before an urgent request arrives.

If you already responded

Stop communicating through the suspicious channel. Contact the relevant bank, card provider, or service using its official contact route and explain what happened. Preserve messages, transaction information, and timestamps. If credentials were disclosed, secure the affected account through the real service. Report fraud through the appropriate official reporting channel, such as the FTC's reporting site in the United States.

Do not feel obliged to prove the message was AI-generated before asking for help. The important issue is the suspected deception and any exposure of money or accounts. Share the method with friends without forwarding live scam links or publishing private details. A calm, repeatable response is more useful than trying to outguess every new impersonation technique.