deepfake

American  

[deep-feyk] / ˈdipˌfeɪk /

Or deep fake

noun

1. a fake, digitally manipulated video or audio file produced by using deep learning, an advanced type of machine learning, and typically featuring a person’s likeness and voice in a situation that did not actually occur.

   Pornographic deepfakes have caused real harm to women whose faces have been superimposed over those of porn actresses.

---

Etymology

Origin of deepfake

First recorded in 2015–20; deep (learning) ( def. ) + fake ¹ ( def. )

Explanation

A deepfake is an altered photo or video that's completely different from the original but still looks totally realistic. A film that appears to show your school principal winning the U.S. Open — it's probably a deepfake. Deepfakes are manipulated recordings that are made using artificial intelligence. Unlike older techniques for creating false images, deepfakes seem so authentic that you're not sure what's true and what isn't. They can be used for entertainment, to bring back long-dead Hollywood stars or swap characters from different movies. But deepfakes are often intended to deliberately mislead people by spreading fake (but real-looking) information. Deepfake was coined in 2017.

Example Sentences

Even ChatGPT-4o, which was used to generate the deepfake images, did not detect all of them, though it performed better than the other models.

From Science Daily • Mar. 26, 2026

Ari Abelson, the co-founder of OpenOrigins, a company that helps identity AI-generated and deepfake content, told Salon that he sees the political maneuverings of these companies as a byproduct of shifting business models.

From Salon • Mar. 23, 2026

Ashley St. Clair, the mother of one of Musk’s children, also sued xAI over deepfake images created by Grok.

From MarketWatch • Mar. 16, 2026

Cowan added: "If you've got those egregious videos, where someone's putting out a hideous deepfake, they're not going to worry about adding that label."

From BBC • Mar. 2, 2026

It called for lessons on using AI to comb transcripts of earnings calls and catch deepfake videos in breaking-news situations.

From The Wall Street Journal • Feb. 5, 2026