AI & deepfakes
When the voice on the phone is fake
Scammers can now clone a voice from a few seconds of audio and fake a face on a video call. It sounds frightening — but the defenses are simple habits, not high tech. Here's what actually works.
Red flags of an AI scam
- Urgency and secrecy — “I’m in trouble, don’t tell anyone, act now.”
- A request for money by gift card, wire, or crypto — the irreversible kinds.
- A call or voicemail that sounds like a loved one but comes from an unknown number.
- A video call where the face looks slightly off, blinks oddly, or audio lags the lips.
- Pressure not to hang up and verify — real family will understand a callback.
Agree on a family “safe word”
Pick a private word or question only your family knows, and use it to verify any urgent call asking for money or help. An AI clone of your voice can’t know your safe word. This single habit defuses the most devastating version of these scams.
Always hang up and call back
If a “relative,” “bank,” or “boss” calls in a panic, hang up and call them back on a number you already have. Scammers rely on keeping you on the line. A 60-second callback breaks the spell.
Slow down — urgency is the weapon
Every AI scam depends on rushing you past your common sense. The moment a message demands instant action and secrecy, treat that as the red flag itself, not the details.
Lock down the raw material
Voice clones and deepfakes are built from your leaked data and public videos. The less of your information is exposed, the harder you are to imitate. Trim your data-broker footprint and keep your accounts breach-free.
Free tools that can help
We're honest about this: no app — including ours — can reliably detect every deepfake, especially over a normal phone call. Detection tech is improving but imperfect. A few free third-party tools can give you a second opinion:
- Hiya Deepfake Voice Detector — a free browser extension that flags likely AI-generated voices in calls and videos.
- DeepFake-o-meter — a free, open-source analyzer from a university lab for images and video.
These are independent tools we don't control or earn from — treat their results as a hint, never a final verdict. Your safe word and a callback are still your best defense.
Shrink what scammers can use against you
Fewer leaks and exposed details mean less raw material for a convincing fake. TrueID helps you check a link, lock down your accounts, and clean up your data trail.
TrueID.Help is a protection toolkit, not an insurance policy, and we make no claim to detect deepfakes. This page is general guidance to help you recognize and avoid AI-powered scams.