What is changing now
The current pattern.
Recent official warnings describe scammers combining AI-generated voices or videos with spoofed websites, government branding and urgent requests. The synthetic media builds confidence, but pressure, secrecy and payment instructions still do most of the manipulation.
How it works
Confidence first, then pressure.
A scammer collects public audio, video, photographs or biographical details about a person. Generative tools can then produce a convincing voice message, live-call effect or video that appears to show that person speaking.
The contact usually supplies an emotional story: a relative is in trouble, an executive needs an urgent transfer, or an official says you are under investigation. The request may move across email, messaging apps, calendar invitations and voice or video calls so that each channel appears to confirm the others.
Visual glitches and unnatural speech can be clues, but they are not a dependable test. High-quality synthetic media may look or sound normal, while a genuine call can lag or distort. Independent contact is more reliable than trying to become a deepfake detector.
Worked example
A realistic pattern, separated into evidence.
This composite example is educational. It does not describe a real person or reproduce a live malicious message.
A familiar voice asks for an unfamiliar payment
- 01The opening
A caller sounds like a relative and says they have lost their phone while travelling. They know a family nickname and mention a real upcoming event, which makes the story feel private and current.
- 02The escalation
The caller says they cannot speak for long, asks for secrecy and needs a transfer to a new account before a deadline. When offered a callback, they say their replacement phone cannot receive calls.
- 03The decision point
The decisive clue is not whether the voice sounds synthetic. It is the attempt to keep verification inside a channel controlled by the caller while introducing urgency, secrecy and a new payment destination.
Warning signs
Reasons to stop and verify.
- A sudden emergency, disciplinary threat or confidential business request
- Instructions to stay on the line, keep the situation secret or avoid colleagues and relatives
- A payment by wire, cryptocurrency, gift card or a newly supplied account
- A new number or messaging account claiming an old identity
- Resistance when you suggest ending the call and contacting the person separately
- A request for identity documents, account credentials or a one-time code
Independent verification
Test the request without using its evidence.
These checks deliberately move the decision away from the person, link, number or account that introduced the request.
- 01
Separate identity from the story
Write down who the caller claims to be and what they want. Do not treat a familiar voice, face, number or biographical detail as independent proof.
- 02
Start a clean contact
Use a saved number, an existing account conversation or a second trusted person. Do not use a callback number or account introduced during the request.
- 03
Verify the event and the payment separately
Confirm both that the emergency exists and that the named person controls the destination account. Verifying only one of those facts is not enough.
- 04
Use a family or team protocol
Agree in advance that urgent money requests require a callback and a second person. A private phrase can help, but it should supplement—not replace—a separate contact route.
Safer response
Move the decision outside the contact.
- 01
Break the channel
End the call or meeting. Do not use a callback number, link or account supplied during the contact.
- 02
Use a known route
Call a saved number, open an existing conversation, or contact a colleague or relative who can verify the story independently.
- 03
Ask a separate question
Use information not available in public posts or the current conversation. Families and teams can agree on a private verification phrase in advance.
- 04
Verify authority claims
Find the organisation’s official public number yourself. Government and law-enforcement branding inside a message is not proof.
If you already acted
Protect access and contact the real provider.
- Contact your bank or payment provider immediately if money was sent.
- Change exposed passwords from a trusted device and end unfamiliar account sessions.
- Tell the person or organisation being impersonated so they can warn others.
- Preserve the original contact details and report the incident to the relevant platform and local authority.
- Be cautious of anyone who later promises recovery for an upfront fee; previous victims are often targeted again.
Sources
Official guidance used for this article.
Stop & Verify summarises these sources for general education. Reporting routes and legal protections vary by location.