As of now, copyright holders haven't caused issues with the app, but Benkhin said 'who the hell knows what's going to happen?' The app allows anyone to use any image, with a majority of the popular shares using copyrighted characters and songs. Using some 'complicated math' that Benkhin does not 'have the exact technical description of' the app processes the image and spits out a cute video or an abomination of cursed excellence. That motion is applied to a model that is then mapped with the images users submit. A performer executes a series of facial, eye, and lip movements along with the song. How does Wombo.ai work?īehind each song is a driving video that the developers make in their studio using motion capture. With just seven people on staff, Wombo.Ai rolled out a beta in January before launching and immediately picking up steam with its odd creations. He first got the idea for Wombo in August of last year, capitalizing on his affinity and love for artificial intelligence technology.

'My whole goal with Wombo is to make it easy for anyone to make a hilarious, shareable, and entertaining piece of content of themselves.'īenkhin spent the previous five years running a website design agency in Toronto before taking the leap and designing an app of his very own. 'One can get in front of a camera and sing a song and do a funny or interesting or entertaining performance, but with AI, anyone can,' CEO of Wombo.ai Ben-Zion Benkhin told Insider.

Using artificial intelligence to transform any picture into a singing video, the app has soared up the iOS and Play Store charts with over 2.2 million downloads and 15 million Wombo videos created. Over the past two weeks, the Wombo.Ai app has spread across social media, creating deepfake copies of figures like Kim Jong-Un or the Mona Lisa lip-syncing to a selection of meme songs.

Use deepfake app