Artificial-intelligence tools have become good enough at fabricating realistic video that the FBI is urging people to confirm the identity of anyone contacting them through a second, independent channel before trusting the message. In a public warning, the bureau described scammers circulating deepfake clips of senior officials to lend credibility to fraud schemes, part of a wider surge in synthetic-media impersonation. The guidance reflects a shift in the threat landscape: seeing a familiar face on screen or hearing a familiar voice can no longer be treated as proof that a request is genuine.
The FBI’s public warning
The bureau issued its alert through the Internet Crime Complaint Center, which publishes public service announcements about emerging online fraud. The July 2026 announcement described a rise in social-engineering schemes that blend impersonation, synthetic media, and phishing to exploit people, including some who had already been victimized once. One documented scheme relied on AI-generated video showing what appeared to be a senior FBI official directing viewers to file complaints through a site built to mimic the bureau’s own, a page whose only working function was a form that harvested the victim’s name, contact details, and reported losses.
The bureau’s core recommendation is straightforward: verify the person or organization through a separate, trusted method rather than the one used to make contact. If a caller claims to be from an agency or company, the recipient can independently look up an official number and call it back, refusing to rely on any contact information the suspicious message supplies. That out-of-band verification is designed to defeat impersonation regardless of how convincing the audio or video appears.
How the deepfake schemes operate
Deepfake fraud takes several forms. In recovery scams, criminals target people who have already lost money, posing as investigators who can retrieve the stolen funds in exchange for a fee or additional personal information, and using fabricated video of officials to appear authoritative. In other schemes, synthetic voice or video mimics a family member, executive, or government agent to pressure a target into transferring money urgently. The techniques exploit trust and urgency, the same psychological levers behind older scams, but with a far more persuasive veneer. The bureau’s Internet Crime Complaint Center collects reports of these incidents and channels them into investigations and further alerts.
The FBI notes that deepfakes may still contain detectable flaws, such as unnatural facial movements, distorted features, mismatched or lagging audio, and inconsistencies in accompanying email addresses, phone numbers, or web addresses. But it cautions that these tells are not always obvious and are steadily becoming harder to spot as the underlying technology improves, which is why the agency stresses verification over visual inspection.
Why a second channel matters
The logic of the FBI’s advice is that authentication should not depend on the medium an attacker controls. A scammer who can fabricate a video can also spoof a caller ID or clone a website, so any single channel can be faked convincingly. By confirming through a separate, independently sourced route, a phone number found on an official site, an in-person check, or a previously established contact, a person breaks the attacker’s control over the interaction. Federal agencies have promoted this out-of-band verification as one of the few defenses that holds up even as synthetic media grows more realistic. Broader guidance from the FBI reinforces that legitimate agencies do not pressure people into immediate payments or demand secrecy, and that unexpected urgency is itself a warning sign.
The wider deepfake threat
The warning about scam videos fits into a larger official concern about synthetic media being used for fraud, disinformation, and manipulation. Security specialists have documented deepfakes deployed against businesses in schemes that impersonate executives to authorize fraudulent transfers, as well as against individuals in romance and investment cons. Coverage from Infosecurity Magazine highlighted the bureau’s specific concern about fabricated videos impersonating its own leadership, a tactic that weaponizes public trust in law enforcement.
For the public, the practical defense the FBI keeps returning to is habit rather than technology: treat unsolicited contact as unverified until confirmed independently, resist pressure to act immediately, and report suspected impersonation through the official complaint center. As AI-generated video approaches the point where flaws are no longer reliably visible, agencies argue that the burden of proof has to shift, and that a familiar face or voice should prompt verification rather than automatic trust.
Adapting habits to a synthetic-media era
The broader lesson security officials draw from the deepfake warnings is that long-standing verification habits have to be updated for a moment when audio and video can no longer be taken at face value. Practices once reserved for high-stakes corporate transactions, such as confirming an unusual request through a pre-agreed code word or a callback to a known number, are increasingly recommended for ordinary interactions with relatives and institutions. Awareness campaigns emphasize that the technology behind convincing fakes is now widely accessible, meaning the defense cannot rest on assuming an attacker lacks the sophistication to produce them. The durable protection, officials repeat, is procedural: build in a verification step that a forged image or cloned voice cannot satisfy.
This article was produced with the assistance of AI and reviewed by Morning Overview editors prior to publication.
More from Morning Overview
