People take their faces off so they can be a deepfake-style marketing clone
Liri can do a lot of work, in many countries, since she took her face off in Hour One, a startup that uses human similarities to create AI -spoken characters that will immediately go on sale and in videos. educational for organizations around the world. It’s about a wave of companies improving the way digital content is created. And this has huge implications for human worker.
Liri does her waitressing and bar work personally, but she has little idea what’s in her digital clones. “It’s really a bit weird to think that my face can be seen in videos or ads for different companies,” he said.
Time One isn’t the only company getting deepfake tech mainstream, use it to mash-up real-time AI-generated footage and video. Others used professional actors to adding life to deepfaked personas. But Hour One doesn’t require any specific expertise. You have to be willing to give rights in your face.
Hour One grows a pool where they are called “characters.” It says there are about 100 of its books to date, with more being added each week. “We have some of the people who are about to become these characters,” said Natalie Monbiot, the company’s chief strategist.
Anyone can apply to become a character. Like a modeling agency, Hour One filters through applicants, who choose what they want from its books. The company targets a wide sample of characters that reflect the age, gender, and background of people in the real world, according to Monbiot. (Today, nearly 80% of its characters are under the age of 50, 70% female, and 25% white.)
To create a character, Hour One will use a 4K high -resolution camera to capture a person talking and make different facial expressions in front of a green screen. And that’s for the man’s part to do. Entering the resulting data into AI software that works in much the same way as deepfake tech, Time One can create an endless amount of capture of the person speaking whatever it wants, in any language.
Time One clients pay the company to use its characters in a promotional or commercial video. They chose a face, uploaded the text they wanted to say it, and played back a video of what a real person would look like delivering that script to a camera. The fastest service uses text-to-speech software to create synthetic sounds, which are combined with the characters ’mouth movements and facial expressions. Hour One also provides a premium service where audio is recorded by professional voice actors. These voices also correspond to the actions of the character in the video. Hour One says it has more than 40 clients, including home and house, e-commerce, digital healthcare, and corporate firms. A major client is Berlitz, an international language school that offers teacher -led video courses for multiple languages.
According to Monbiot, Berlitz wants to increase the number of videos it offers but struggles to do so using real human artists. They have to have production crews doing the same set up with the same actor on a regular basis, he says: “They find that it’s never really nutritious. We’re talking about thousands of videos.”
Berlitz now works at Hour One to produce hundreds of videos in minutes. “We replaced the studio,” Monbiot said. “Someone doesn’t have to waste their time filming.”