• D.C.
  • BXL
  • Lagos
  • Dubai
  • Beijing
  • SG
rotating globe
  • D.C.
  • BXL
  • Lagos
Semafor Logo
  • Dubai
  • Beijing
  • SG


Updated May 9, 2023, 10:19am EDT
techEast Asia

As AI explodes on TikTok, its Chinese counterpart requires labels for AI-generated content

Douyin.
Oriental Image via Reuters
TweetEmailWhatsapp

Sign up for Semafor Technology: What’s next in the new era of tech. Read it now.

Title icon

The News

Music, photos, and videos created by artificial intelligence have exploded in popularity on TikTok in recent months, including clones of singers’ voices or images of the Harry Potter cast reimagined as Star Wars characters.

This AI boom prompted the Chinese-owned short video app to put out new rules in March requiring users to disclose if they post any AI-generated content that shows “realistic scenes.”

Now, Douyin, the Chinese version of TikTok that is also owned by the China-based ByteDance, is going a step further following new Beijing regulations — requiring creators to label anything created with AI.

AD
Title icon

Know More

Douyin disclosed its new rules on WeChat on Tuesday. It said users should prominently mark AI content and be held responsible for the consequences of posting AI images and video.

It also said that anyone who uses a virtual avatar on Douyin should register with a real name, and that using AI to spread misinformation or rumors is not allowed.

Title icon

Step Back

Amid the recent AI explosion, Chinese officials have kept a close eye on the technology and in January rolled out new rules aimed at curbing the spread of misinformation and deepfakes posted on social media. In February government regulators also warned the public to be wary of “rumors” generated by ChatGPT.

AD

On Sunday, police in northwestern China said they arrested a man who allegedly admitted to using ChatGPT to create fake news and spread it online. It appeared to be the first time Beijing detained a citizen for using the chatbot.

Title icon

The View From TikTok

TikTok’s newest rules related to AI went into effect on April 21, covering “synthetic or manipulated media that shows realistic scenes.”

Under the rules, any AI that depicts a “real private figure” is not allowed.

AD

There’s more leeway given to public figures over the age of 18, but “we do not want them to be the subject of abuse, or for people to be misled about political or financial issues,” the policy states.

Deepfakes of artistic content, for instance, are allowed, “such as a video showing a celebrity doing a popular TikTok dance.” But deepfakes of public figures can’t be used for endorsements or to spread hate speech.

Title icon

Notable

  • TikTok is working on a tool that will allow creators to more easily label when something in their video is made with AI, The Information reported last week.
  • The question of AI labels goes beyond just videos on social media. A proposed U.S. Congress bill would mandate that political ads disclose when they use AI-generated content, after the Republican National Committee recently debuted an ad created with AI, The Washington Post reported.

— with Diego Mendoza

Semafor Logo
AD