China’s AI news anchor can only answer preset questions with propagandistic answers

Meet the latest artificial intelligence (AI) powered broadcast journalist recruited by China’s state-controlled newspaper.

The virtual young woman, named “Ren Xiaorong,” claims to be able to answer questions and deliver news broadcasts 24 hours a day, 365 days a year.

But there’s a catch — the avatar can only answer pre-set questions, and the answers it gives strongly promote the line of the Communist Party of China (CPC) Central Committee.

The questions relate to the “Two Sessions” political conference that ended on Monday, at which delegates from the country discussed and passed laws.

Developed by the CCP’s official news outlet, People’s Daily, Ren says she has the skills of “thousands of newscasters.”

She was revealed in a video on Weibo, a social networking platform in China, dressed smartly in a black blazer and standing in front of a city skyline.

The digital news anchor says, “Hello, my name is Ren Xiaorong. I’m a digital AI anchor who just joined People’s Daily.

“Thousands of new anchors have taught me their professional skills.

“I will report news for the whole year, 24 hours a day, 365 days a year, around the clock, without a break.

“Whether it’s on the news sites or in the studio, you’ll always see me.

“Every conversation, every feedback you give will only make me smarter.”

Despite these claims, when you open the app to interact with Ren, you can only ask her to talk about preset topics related to the two sessions.

These include ‘High Technology’, ‘Tax and Fee Reduction’, ‘Employment’ and ‘The Belt and Road Initiative’.

All of her answers are consistent with the CCP’s ideology, and the user is unable to make her deviate from the script by asking her her own question.

Reactions to her on Weibo have been mostly positive, with comments ranging from “wow!”, “amazing,” “so real,” and “very sci-fi.”

But the popular microblogging site is under surveillance to comply with China’s strict censorship regime.

Platforms across the country are quick and regular in deleting articles, from complaints about COVID lockdowns to cryptic criticism of politics.

On Twitter, which is not accessible in China, the reaction is less positive.

One user said: “This looks so bad. In the meantime we have developed deep fakes, AI dialogues and AI voices.”

Ren is not the first AI-generated presenter in China as he has been appearing online and on TV since 2018.

The first three were modeled after three existing reporters, two men and one woman, for the Xinhua News Agency.

Next came the robot host Jiang Lailai, who presented an entertainment program about new technologies.

It wasn’t long before Xinhua recruited another digital anchor, Xin Xiaomeng, who was modeled after another journalist and also shared news of the two sessions in 2020.

However, these all look relatively dated now, especially when compared to ChatGPT’s ever-expanding capabilities.

That doesn’t mean China has dropped out of the AI ​​chatbot race, however, as today it unveiled its own competitor, Ernie, developed by tech company Baidu.

However, it disappointed investors with its use of pre-recorded video and lack of a public launch, causing its shares to plummet.

It’s only available for a first batch of users to test, and even Baidu’s CEO Robin Li admitted it was only introduced “because the market demands it.”