Published on November 9th, 2018 | by Emergent Enterprise0
China Created What it Claims is the First AI News Anchor
Photo: Xinhua’s English-language artificial-intelligence anchor. New China TV/YouTube
- China’s state press agency has developed what it calls “AI news anchors,” avatars of real-life news presenters that read out news as it is typed.
- It developed the anchors with the Chinese search-engine giant Sogou.
- No details were given as to how the anchors were made, and one expert said they fell into the “uncanny valley,” in which avatars have an unsettling resemblance to humans.
China’s state-run press agency, Xinhua, has unveiled what it claims are the world’s first news anchors generated by artificial intelligence.
Xinhua revealed two virtual anchors at the World Internet Conference on Thursday. Both were modeled on real presenters, with one who speaks Chinese and another who speaks English.
“AI anchors have officially become members of the Xinhua News Agency reporting team,” Xinhua told the South China Morning Post. “They will work with other anchors to bring you authoritative, timely, and accurate news information in both Chinese and English.”
In a post, Xinhua said the generated anchors could work “24 hours a day” on its website and various social-media platforms, “reducing news production costs and improving efficiency.”
Xinhua developed the virtual anchors with Sogou, China’s second-biggest search engine. No details were given about how they were made.
Though Xinhua presents the avatars as independently learning from “live broadcasting videos,” the avatars do not appear to rely on true artificial intelligence, as they simply read text written by humans.
“I will work tirelessly to keep you informed as texts will be typed into my system uninterrupted,” the English-speaking anchor says in his first video, using a synthesized voice.
You can watch the first appearance of the English-language “AI news anchor,” who is modeled on the real presenter Zhang Zhao, here:
Convincing though it might seem at first glance, the movement of the mouth is clearly edited, the facial expression seems limited, and the voice is also highly robotic.
The Oxford computer-science professor Michael Wooldridge told the BBC that the anchor fell into the “uncanny valley,” in which which avatars or objects that closely but do not fully resemble humans make observers more uncomfortable than ones that are more obviously artificial.
“As an AI news anchor under development, I know that there is a lot for me to improve,” the virtual anchor says as it signs off its report.”