In the past few days, a new software package called Deep-Live-Cam has been making waves on social media, drawing attention for its ability to create real-time deepfakes.
A hot potato: As deepfake technology continues to evolve, the potential for misuse grows. While current tools still require users to mimic mannerisms, voice, and other details, advancements in voice cloning and video synthesis could make creating digital doppelgängers in real-time even more convincing.
In the past few days, a new software package called Deep-Live-Cam has been making waves on social media, drawing attention for its ability to create real-time deepfakes with incredible ease. The software takes a single photo of a person and applies their face to a live webcam feed, tracking the person’s pose, lighting, and expressions on the webcam. While the results are not flawless, the technology’s rapid advancement underscores how much easier it has become to deceive others with AI.
Ars Technica notes that the Deep-Live-Cam project has been in development since late last year, but it has recently attained viral attention after example videos began circulating online.
