What Is a Deepfake Phone Call?

The Definition

A deepfake phone call is a phone call in which the caller's voice is synthesized or replaced by AI — using voice cloning technology — to impersonate a real person. The AI-generated voice is acoustically indistinguishable from the genuine speaker, speaking in real time during a live call.

The term combines "deep learning" (the AI technology powering voice synthesis) with "fake" — a voice that is artificially generated rather than produced by a real human speaker. Deepfake phone calls are the primary delivery mechanism for AI voice cloning fraud.

Unlike older forms of voice fraud that relied on impersonators trained to mimic specific people, deepfake phone calls use neural networks to generate the target's voice from a short audio sample. Any voice that has ever been recorded — even briefly — can be cloned.

How a Deepfake Phone Call Is Made

Creating a deepfake phone call requires three components: a voice sample, a cloning tool, and real-time voice conversion software. The process takes minutes:

1. Harvest the voice sample — The attacker finds 3+ seconds of audio from the target speaker. This could be a social media video, a podcast clip, a YouTube interview, a voicemail, or any other recorded audio.
2. Generate the voice clone — The audio is fed into an AI voice cloning model, which extracts the speaker's biometric voice characteristics and creates a synthesizer that can produce new speech in that voice.
3. Make the call with real-time voice conversion — The attacker uses real-time voice conversion software during the call. They speak normally; the software converts their voice to the cloned voice in real time, with latency under 100 milliseconds. The recipient hears the cloned voice.

This entire workflow is available through consumer applications, many of which are free or cost a few dollars per month. No specialized technical knowledge is required. Any motivated person can make a deepfake phone call today.

Common Deepfake Phone Call Scams

The Grandparent Scam

The most emotionally devastating deepfake call scam. A criminal clones the voice of a grandchild (sourced from social media) and calls an elderly grandparent claiming to be in trouble — arrested, in a car accident, in the hospital — and needing money immediately. The AI voice clone makes the call convincingly real. Elderly individuals are often asked to wire money, purchase gift cards, or make a cash withdrawal. Read our full guide to the grandparent voice cloning scam.

CEO Fraud / Business Email Compromise by Voice

An attacker clones the voice of a company's CEO or senior executive — using audio from earnings calls, conference presentations, or media appearances — and calls a finance employee to authorize an urgent wire transfer. The employee hears what sounds like their boss's voice and complies. Individual CEO fraud via voice attacks have yielded losses exceeding $25 million in documented cases.

Bank and Financial Institution Impersonation

A caller impersonates a bank representative, IRS agent, or other financial authority, requesting account details or immediate payment to resolve a fabricated emergency. While many such calls use human fraudsters, AI voice cloning is increasingly used to impersonate specific known representatives whose voices may be available online.

Family Emergency Scams

Variations on the grandparent scam targeting a broader audience — cloning the voice of any family member (a spouse, parent, sibling, or child) to manufacture an emergency requiring immediate financial assistance from the recipient.

Why Deepfake Phone Calls Are So Effective

The effectiveness of deepfake phone call scams comes from a psychological reality: we trust people we recognize. When you hear a voice that sounds exactly like your child, your boss, or your bank manager, your guard drops immediately. The emotional response overrides skepticism.

Deepfake phone calls exploit several cognitive vulnerabilities:

• Voice familiarity — we use voice as a primary identity signal; hearing a familiar voice triggers automatic trust
• Urgency and emotion — the manufactured emergency activates a stress response that inhibits critical thinking
• Authority — hearing a boss or official authority figure short-circuits normal verification instincts
• Isolation — "don't tell anyone yet" instructions prevent the victim from consulting others who might be skeptical

How VeriCall Detects Deepfake Phone Calls

VeriCall solves the deepfake phone call problem by performing what your ears cannot: real-time biometric speaker verification — comparing the voice on the call against a stored biometric model of the real person's voice, on your device, in under one second.

When a contact calls you, VeriCall immediately compares the incoming voice to the voiceprint it has built from real, genuine calls with that person. A live confidence score tells you: is this the same person whose voice I know, or is this a deepfake?

If VeriCall shows VOICE VERIFIED — the voice matches the real person's biometric. If VeriCall shows a red alert — the voice does not match. Hang up.

All of this happens on your device. No audio is transmitted to any server. No voiceprints leave your phone. The detection happens in real time, before you've committed to any action.