A deepfake is a fake video, photo or voice recording made with AI that looks or sounds real. Most often it shows a real person saying or doing something they never said or did.
Think of a ventriloquist’s dummy, except the dummy now has your friend’s face and voice. Someone else is pulling the strings and choosing the words, but it looks and sounds like your friend talking.
In a sentence: “That clip of the mayor ‘admitting’ to fraud turned out to be a deepfake; his voice and face had been copied from old interviews.”
What counts as a deepfake
Two things make a deepfake: it copies something real, and it could fool you. The European Union’s AI law, the EU AI Act, gives a legal definition. It covers made-up or doctored images, sound and video of real people, and also of real places, objects or events.1
So an obviously cartoonish AI picture is not a deepfake. A realistic clip of a politician giving a speech they never gave is. In everyday talk, “deepfake” usually means the second kind: a real person, faked.
Where the word comes from
The word was coined in 2017. It joins “deep”, from deep learning (the AI method behind it), with “fake”.2
That December the tech site Vice reported on fake pornographic videos of celebrities made by a Reddit user who called himself “deepfakes”, using AI to put their faces onto other people’s bodies.3
A simple example
The most common one you might meet is a cloned voice on the phone. A scammer copies a family member’s voice from a short clip online, then calls in a panic asking for money. The US Federal Trade Commission warned about this kind of scheme in 2023.4
Video can be faked too. In 2024, an employee at the engineering firm Arup joined a video call in which the senior colleagues he saw were all fakes, and sent the money they asked for.5
Why it matters to you
Seeing and hearing are no longer proof. You cannot rely on spotting a deepfake by eye or ear. The defences that work do not depend on it:
- Agree a family code word and ask for it in any emergency call.
- Hang up and call back on a number you already have saved.
- Treat urgency and secrecy as the warning, especially with requests for gift cards, crypto or wire transfers.
Deepfake self-defence, the full explainer on this site, walks through each step.
Some laws now require labels. The EU AI Act says people who use AI to make deepfakes must say so, with lighter rules for art, satire and fiction.6 That rule applies in the European Union. Other countries have their own rules; provenance and content labels covers the labelling side.
Often confused with
| Term | How it differs |
|---|---|
| Any AI image | Differs: a deepfake imitates something real and could be taken as genuine. A fantasy picture is just AI art. |
| Photo editing | Differs: fixing red-eye or cropping is ordinary editing. Swapping a face or putting words in someone’s mouth is a deepfake. |
| AI slop | Differs: slop is low-effort AI content made in bulk; it need not copy anyone. See AI slop meaning. |
Questions people ask
Are deepfakes illegal?
It depends where you live and what the deepfake is used for; the rules differ from country to country. In the European Union, the AI Act also says anyone using AI to make a deepfake must disclose it.6
How can you tell if something is a deepfake?
Often you cannot, by eye or ear, so check the story instead of the picture or voice. For a panicked call from a relative, the US Federal Trade Commission’s advice is to hang up and ring them back on a number you know.7
Is a deepfake always a video?
No. A faked photo or a cloned voice counts too. The EU’s legal definition covers images, audio and video.1 See voice cloning meaning for the audio kind.
Learn more on this site
- Deepfake self-defence — the full explainer: code words, callbacks and the scam script.
- How AI fraud actually works — seven common scams, including faked calls.
- The household guide — protecting your family from AI scams.
- The AI glossary and AI for beginners — the basics, in plain words.
Related on this site: Voice cloning meaning · How to Make a Talking Photo or Lip-Sync Video With AI, With Consent · Is This Image or Video Real? · Deepfake vs AI-Generated Image · What Is Lip Sync in AI?
QUICK CHECK · 3 QUESTIONS
Pick an answer to see if you got it. Nothing is recorded.
1. What makes something a deepfake?
- Any picture made with AI, even a cartoon
- It copies something real, like a person, and could fool you
- Any photo that has been cropped or edited
Show the answer
Answer: It copies something real, like a person, and could fool you. A deepfake copies something real and could be taken as genuine; an obviously cartoonish AI picture is not one.
2. What does the page say about spotting a deepfake by eye or ear?
- It is easy if you look closely at the eyes
- Only experts with special equipment can do it
- You cannot rely on it, so check the story instead
Show the answer
Answer: You cannot rely on it, so check the story instead. Seeing and hearing are no longer proof, so the defences that work do not depend on spotting the fake.
3. You get a panicked call from a relative asking for money. What does the page suggest?
- Ask for your family code word, or hang up and call back on a number you know
- Send the money quickly in case it is real
- Ask them to send a photo to prove who they are
Show the answer
Answer: Ask for your family code word, or hang up and call back on a number you know. The page suggests agreeing a family code word and calling back on a number you already have, and treating urgency and secrecy as the warning.
Related terms: AI ethics · AI governance · Explainable AI (XAI) · Guardrails · Jailbreak · Watermarking
- 1EU AI Act, Article 3(60), via artificialintelligenceact.eu, Article 3: Definitions, Regulation (EU) 2024/1689, read at source 1 Oct 2026“‘deep fake’ means AI-generated or manipulated image, audio or video content that resembles existing persons, objects, places, entities or events and would falsely appear to a person to be authentic or truthful”. ↑a“‘deep fake’ means AI-generated or manipulated image, audio or video content that resembles existing persons, objects, places, entities or events and would falsely appear to a person to be authentic or truthful”. ↑b
- 2Dictionary.com, deepfake, dictionary entry, read at source 12 Oct 2026“First recorded in 2015–20; deep (learning) + fake”; “Deepfake was coined in 2017.” ↑
- 3Vice (Samantha Cole), report on AI-made fake videos by a Reddit user called ‘deepfakes’, 11 Dec 2017, read at source 12 Oct 2026the videos were made by “a Redditor who goes by the name ‘deepfakes’”. ↑
- 4US Federal Trade Commission, Scammers use AI to enhance their family emergency schemes, consumer alert, 20 Mar 2023, read at source 1 Oct 2026“All he needs is a short audio clip of your family member’s voice — which he could get from content posted online — and a voice-cloning program.” ↑
- 5CNN Business, report on the Arup deepfake call, 16 May 2024, read at source 17 Sep 2026 (as cited on this site’s fraud page): the employee met people “he believed were the chief financial officer and other members of staff, but all of whom turned out to be deepfake re-creations.”↑
- 6EU AI Act, Article 50(4), via artificialintelligenceact.eu, Article 50, read at source 1 Oct 2026“Deployers of an AI system that generates or manipulates image, audio or video content constituting a deep fake shall disclose that the content has been artificially generated or manipulated.” ↑a“Deployers of an AI system that generates or manipulates image, audio or video content constituting a deep fake shall disclose that the content has been artificially generated or manipulated.” ↑b
- 7US Federal Trade Commission, Scammers use AI to enhance their family emergency schemes, consumer alert, 20 Mar 2023, read at source 12 Oct 2026“Don’t trust the voice.” “Call the person who supposedly contacted you and verify the story.” “Use a phone number you know is theirs.” ↑
EU AI Act (Regulation (EU) 2024/1689), Article 3: Definitions and Article 50, via artificialintelligenceact.eu, read at source 1 Oct 2026.
The comparison on this page (the ventriloquist’s dummy) is this site’s own way of explaining, not a claim from these sources.
Free printable AI cheat sheets, and AI news every week.
Sixteen one-page AI guides and checklists in one free PDF, no sign-up. And if you want to keep up, the newsletter covers what changed in AI each week, in plain English, with the sources.
SUBSCRIBE FREE ↗ GET THE CHEAT SHEET PACK