deepfake
Les deepfakes sont de plus en plus difficiles à détecter.
Deepfakes are becoming harder to detect.
Manipulated media (image, video, audio) created using artificial intelligence to convincingly replace or mimic someone's likeness or voice.
Les deepfakes sont souvent utilisés pour diffuser de la désinformation.
Deepfakes are often used to spread disinformation.
Cette vidéo est probablement un deepfake, car les expressions faciales semblent artificielles.
This video is likely a deepfake because the facial expressions seem artificial.
The term is a portmanteau of 'deep learning' and 'fake'.
Collocations
Pro Tip
Avoiding confusion
Deepfake refers specifically to AI-generated media, not just any edited or altered content. For example, a Photoshopped image is not a deepfake unless it uses AI techniques like neural networks.
Gold Rule
Context matters
Use 'deepfake' when discussing AI-generated media, especially in technology, media, or legal contexts. Avoid using it for simple edits or traditional forgeries.
Word Origin
Blend of 'deep learning' (artificial intelligence technique) and 'fake' (counterfeit). First recorded in 2017 in the context of AI-generated media.
Usage Notes
The term is widely recognized in French but remains a loanword. It is often used in discussions about misinformation, cybersecurity, and ethical concerns in AI.