With a state-run media like China’s, there’s already some concern that newscasters are little more than puppets. After an AI news anchor debuted at the World Internet Conference in China this week, we’re one step closer to that reality.
The anchor was created in a partnership between Xinhua News Agency, China’s official state-run media outlet, and sogou.com, a Chinese search engine company. The Chinese news, of course, is thrilled and impressed, claiming that the character “can read texts as naturally as a professional news anchor.” Two versions of the AI anchors are now available on Xinhua through their apps, WeChat account, and online news channel.
The AI anchor in this video was modelled after one of the network’s existing anchors and, as Xinhua points out, he can work 24 hours a day—as long as a human inputs things for him to say. Although the voice is a little robotic and the movement of his mouth can be jumpy, it’s actually quite convincing.
And scary. If this new AI anchor caught on, you could feed as much fake news through the newscaster’s mouth as and whenever you wanted. He’s not going to fight you. There’s also the possibility of completely faked original content. These videos, called DeepFakes, show people saying things they never actually said. In one example this past spring, Buzzfeed News worked with actor and director Jordan Peele to make a fake video of President Barack Obama weighing in on things like Black Panther.
With companies like Photoshop working on a technology to synthesize audio, and Canadian startup Lyrebird, which claims it can convincingly fake someone’s voice by using a minute of actual recordings, the technology is just going to get better. And our ability to trust what we see is going to get much much worse.