- Wombo.ai allows users to transform any image into a singing face.
- CEO Ben-Zion Benkhin transitioned from making websites to creating faces with artificial intelligence.
- The app has already been downloaded over two million since launching two weeks ago.
- Visit Insider’s homepage for more stories.
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. 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.
“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. “My whole goal with Wombo is to make it easy for anyone to make a hilarious, shareable, and entertaining piece of content of themselves.”
Benkhin spent the previous five years running a website design agency in Toronto before taking the leap and designing an app of his very own. He first got the idea for Wombo in August of last year, capitalizing on his affinity and love for artificial intelligence technology. 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.
How does Wombo.ai work?
Behind each song is a driving video that the developers make in their studio using motion capture. A performer executes a series of facial, eye, and lip movements along with the song. That motion is applied to a model that is then mapped with the images users submit. 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.
-Kane⨂ (@FullmetalKaneru) March 10, 2021
The app allows anyone to use any image, with a majority of the popular shares using copyrighted characters and songs. As of now, copyright holders haven’t caused issues with the app, but Benkhin said “who the hell knows what’s going to happen?”
There are currently 15 different tracks you can use on the app, ranging from Rick Astley’s “Never Gonna Give You Up” to internet classics like “Numa Numa.” According to Benkhin, they are still working on “the rights issue” but they do have “arrangements with a few different artists.” Moving ahead, Benkhin hopes that musicians will want to bring their music over to the “because it’ll give them an entirely new way to engage with their audience in a viral and funny way.”
Is Wombo.ai safe?
-PB | ⚠️ Suggestive Warning ⚠️ (@PB_sfw) March 11, 2021
Users can pay for a premium version of the app for $US4.49 ($AU6) a month or $US29.99 ($AU41) a year, giving them faster processing and no ads, or can use the app for free and get served ads. In the future, Wombo hopes to collaborate with musicians or advertisers who can benefit from the cross promotion.
Wombo uses artificial intelligence to create a deepfake, or a reconfigured video close to reality. Deepfake technology has improved drastically over the years and has become accessible to regular people, albeit in a limited capacity. MyHeritage, an ancestry tracking service, released their Deep Nostalgia tool that allows users to animate pictures of deceased relatives and a visual effect artist went viral with a faux Tom Cruise.
“These deepfakes aren’t going anywhere anytime soon,” Benkhin notes. “Over time the public is going to warm up to it…I think we’re only beginning to explore the applications of this technology.”
A deepfake can potentially cause harm, with unaware onlookers potentially confusing realstic enough looking fakes for the real thing. According to Benkhin, Wombo avoids that issue in two ways: the app creates an image that “looks realistic but doesn’t look real” and the limited selection of songs only makes it so “the worst thing you can do is make someone sing.”
The future of Wombo is dependent on if the app can use this viral momentum to pivot from a gimmick or fad into something internet users consistently use. There are social plans for the app that will allow users to create custom profiles and share their faces with friends that can then be transformed.
“It’ll be a place to enjoy and discover new music,” Benkhin said. “And overall, it’ll be the app you go to when you want to laugh and smile.”