Deepfake AI Scams: What They Are and How to Spot Them

In early 2024, a finance worker in Hong Kong sat through a video call with what looked like the company's CFO and several senior colleagues. The faces were familiar. The voices matched. The request — a secret $25 million transfer — sounded urgent and official.

It was all fake. Every person on the call was a deepfake. The money disappeared.

That case was not a one-off. In 2025 and 2026, AI voice clones and deepfake videos have become standard tools for scammers. The technology is cheap, fast, and good enough to fool people under pressure. The question is no longer "Can scammers fake a voice?" — it is "How do I spot it before I act?"

This guide explains what deepfake AI scams look like, where they show up, and the verification habits that stop them.

What deepfake AI scams look like

A deepfake scam has one goal: make you believe someone you trust is asking you to do something. The three most common forms are:

• AI voice-clone calls — a "grandchild" calls in tears asking for bail money, a "CEO" demands an urgent wire transfer, or a "romantic partner" asks for emergency funds.
• Deepfake video calls — synthetic faces and voices on Zoom, Skype, WhatsApp, or Telegram video calls, often used in business-email-compromise (BEC) follow-ups.
• Fake celebrity or influencer endorsements — AI-generated videos of Elon Musk, a local news anchor, or a trusted creator promoting a crypto platform or investment scheme.

The common thread is urgency plus trust. The scammer does not need perfect technology — they only need you to act before you think.

Why they work so well

AI tools that used to require Hollywood budgets now run on a laptop. A scammer can:

• Clone a voice from 3–10 seconds of audio scraped from TikTok, YouTube, a podcast, or a voicemail greeting.
• Generate a convincing talking-head video from a single photo using free or subscription apps.
• Translate and lip-sync the fake into multiple languages, making cross-border scams easier.

Because the fake is built from real data, the result feels personal. The voice sounds like your son. The face looks like your boss. That emotional hit is what bypasses skepticism.

Red flags that expose a deepfake scam

No single sign proves a call is fake, but several together should make you pause:

• Urgent money request — bail, hospital bill, customs fee, locked account, missed tax payment, or "time-sensitive investment."
• Secrecy — "Don't tell Mom," "This is confidential," or "Do it now and explain later."
• Strange pauses or mismatched lip sync — the audio and video do not quite line up, or responses feel slightly delayed.
• Flat or off emotions — the voice sounds right but lacks the natural warmth, sarcasm, or stress of the real person.
• Weird visual artifacts — blurred hairline, flickering around the jaw, unnatural blinking, or eyes that do not track light consistently.
• Refusal to verify — the caller resists a simple test like calling back, turning profile, or answering a question only the real person would know.

If you notice any of these, treat the call as unverified until proven otherwise.

How to verify before you act

The safest response is to break contact and re-establish it through a known channel:

1. Hang up. Do not use any callback number the caller provides.
2. Call the person directly on a number you already have saved — not one from the suspicious message.
3. Use a pre-arranged safe word with family members for emergency calls.
4. For business wires, confirm out-of-band through a second channel (Slack, in-person, or a known phone number) before any transfer.
5. Ask the video caller to perform a live action — turn fully sideways, cover one eye, wave slowly, or hold up a specific object. Many deepfakes glitch on profile views or partial face occlusion.
6. Search the identifier — phone number, email, wallet, or website — on scamers.org to see if others have reported the same pitch.

These steps take two minutes. A deepfake scam relies on you skipping them.

What to do if you are targeted

If you realize a call or video is a deepfake scam attempt:

• Do not send money or share codes, passwords, or 2FA tokens.
• Document everything — screenshots, call recordings if legal in your jurisdiction, phone numbers, wallet addresses, and the exact story used.
• Report to local police or the relevant cyber-crime agency (IC3 in the US, Action Fraud in the UK, etc.).
• [Report a scammer](/report) on scamers.org so the phone number, email, wallet, or website becomes searchable for the next target.
• Warn the person being impersonated — they may need to alert contacts or lock down social accounts.

If you already sent money, contact your bank, card issuer, or crypto exchange immediately. Speed matters: wire reversals and crypto freezes are sometimes possible within the first hours.

The bigger picture: AI is not the scam, trust is

Deepfake technology will keep improving. The defense is not to become a forensic expert — it is to make verification a habit. Any unexpected request for money or sensitive data, especially one that arrives by voice or video, deserves a second-channel check.

For more on voice-specific scams, read our full AI voice scams guide. And if you encounter a deepfake scam attempt, help others avoid it by using report a scammer.

The bottom line

Deepfake AI scams are effective because they weaponize trust and urgency. The best protection is a simple rule: if someone you know suddenly asks for money by phone or video, verify through a separate channel before doing anything. No exceptions. No embarrassment. Two minutes of caution can save thousands of dollars and a lot of regret.