What is a deepfake?
Definition
A deepfake is an image, video or audio clip created or altered with artificial intelligence to make it look or sound like a real person said or did something they never did.
- Category
- Attacks
- Also called
- deep fake, synthetic media, AI-generated fake
- Last updated
The FBI describes deepfakes as a wide category of AI- or machine-learning-generated or altered digital media, including images, video, audio and text. They can show the alteration or impersonation of a person's identity, so it looks like they said or did things they never did.
How deepfakes are made
Software learns from real photos, video or recordings of a person, then produces new content that mimics their face or voice. It no longer takes much expertise or many samples, which is why the volume of fakes has grown.
How they're used
Many deepfakes are harmless or satirical. The harmful uses are the ones that make news:
- Fraud and impersonation. A cloned voice that sounds like a relative asking for money, or a fake video of an executive or official.
- Non-consensual intimate images. Fake explicit images of real people. The FBI has charged people under the TAKE IT DOWN Act for publishing them.
- Fake endorsements and scams. Fake videos of public figures promoting a scheme. A 2026 FBI notice says scammers used AI-generated videos of senior FBI leaders to promote a spoofed IC3 complaint site.
A typical case: you get a call in what sounds like your son's voice saying he's in trouble and needs money wired now. The voice was cloned from a short clip posted online.
Is it illegal?
The FBI says creating or sharing synthetic content isn't illegal in itself, and much of it is protected speech. It becomes a crime when used for things like fraud, extortion or non-consensual intimate imagery. Laws differ by country and state.
Deepfake vs. cheapfake
A deepfake is made with AI. A "cheapfake" is a real clip edited with simple tools, such as slowed down or cut out of context. Both can mislead.
How to protect yourself
- 1
Verify unexpected requests another way
If a voice or video asks for money or sensitive information, hang up and contact the person on a number you already know. Agree on a family code word for emergencies.
- 2
Look for signs of manipulation
Detection tips from FBI-linked materials include blurry or distorted facial features, unnatural blinking, audio out of sync with the video, and odd lighting or shadows. These cues are not reliable, as fakes keep improving.
- 3
Limit the material available to copy
Fakes need source images and audio. Set social profiles to private and think about what you post publicly. Serus finds where your personal details are listed online, including people-search sites, and removes them.
- 4
Be skeptical of too-good or too-shocking content
Check the source and look for coverage from established outlets before sharing.
- 5
Report abuse
Report scams at ic3.gov and ReportFraud.ftc.gov. Report fake or intimate content to the platform that hosts it and ask for it to be taken down.
See where your personal data is exposed
Serus finds it on data brokers and people-search sites and removes it.
Frequently asked
Are deepfakes illegal?
Making one isn't automatically illegal, according to the FBI. Using one to commit fraud, extort someone or publish non-consensual intimate images can be, and some places have laws aimed directly at them.
How can you tell if something is a deepfake?
Look for distorted faces, odd blinking, mismatched lip sync and strange lighting. Better fakes won't show these, so verify through another channel.
Can someone make a deepfake of me?
Yes, if there are enough photos, video or audio of you. Keeping public material limited makes it harder, but can't prevent it entirely.
What should I do if I'm the target of a deepfake?
Save evidence, report it to the platform and to the FBI at ic3.gov or local police, and ask for removal. For intimate images, seek support and report promptly.
Sources
Something wrong or out of date? Tell us and we'll re-check it.