Snap reportedly acquires a deepfake startup

Snap has purchased AI Factory, an image and video recognition startup, according to Variety on Friday. Snapchat reportedly used AI Factory’s technology to launch its new Cameos feature, which allows users to insert selfies into a scene to send as a looping video and raises concerns about the possibility of creating deepfakes.

Deepfakes, video forgeries that make people appear to be doing or saying things they didn’t, are the moving-picture equivalent of bogus images created with programs like Photoshop. Deepfake software has made manipulated videos accessible and increasingly harder to detect as fake.

Snapchat’s Cameos feature — located inside the chat area below Bitmoji and above emoji selections — lets you take a selfie and choose a gender. Snapchat then stores the selfie and manipulates your features to make different facial expressions like smiling, blowing kisses and opening your mouth.

Your manipulated selfie is then inserted on top of things like puppies and bunnies, but also onto other people — like someone in a bathtub, at a hair salon, in a classroom, on a movie set, driving a car, sitting on an airplane and kissing the camera while wearing a skintight dress.

You can clear your Cameos selfie in Snapchat’s settings.

A Ukrainian tech publication reported Snap’s acquisition of AI Factory to be worth $166 million, Variety said. Snap confirmed the acquisition in an email to TechNews Friday afternoon, but did not reveal the terms.

Congress is investigating deepfakes following the appearance of doctored videos of US House Speaker Nancy Pelosi and amid fears that deepfakes could escalate the fake news campaign during the 2020 US presidential race. Social media companies like Twitter and Facebook are also coming under pressure to find ways to more quickly detect and remove deepfakes from their platforms, along with abusive content, terror-related content, misinformation and fake news in the lead-up to the election.

In October, Snapchat reported having 210 million daily users.