Advertisement
Singapore markets closed
  • Straits Times Index

    3,224.01
    -27.70 (-0.85%)
     
  • Nikkei

    40,369.44
    +201.37 (+0.50%)
     
  • Hang Seng

    16,541.42
    +148.58 (+0.91%)
     
  • FTSE 100

    7,952.62
    +20.64 (+0.26%)
     
  • Bitcoin USD

    70,336.78
    -561.68 (-0.79%)
     
  • CMC Crypto 200

    885.54
    0.00 (0.00%)
     
  • S&P 500

    5,254.35
    +5.86 (+0.11%)
     
  • Dow

    39,807.40
    +47.29 (+0.12%)
     
  • Nasdaq

    16,379.46
    -20.06 (-0.12%)
     
  • Gold

    2,254.80
    +16.40 (+0.73%)
     
  • Crude Oil

    83.11
    -0.06 (-0.07%)
     
  • 10-Yr Bond

    4.2060
    +0.0100 (+0.24%)
     
  • FTSE Bursa Malaysia

    1,536.07
    +5.47 (+0.36%)
     
  • Jakarta Composite Index

    7,288.81
    -21.28 (-0.29%)
     
  • PSE Index

    6,903.53
    +5.36 (+0.08%)
     

YouTube says that an error caused comments critical of China's government to auto-delete

YouTube responded to reports that it is automatically deleting comments criticizing the Chinese government, explaining that the seeming censorship is actually an error in its automated moderation systems.

As the Verge reported, comments on the platform using the Chinese phrases for "communist bandit" or "50-cent party" — two terms tied to criticisms of the Chinese Communist Party — were taken down almost instantly, even if those comments were positive. The latter term (五毛 or "wumao dang") refers to China's censorship efforts, particularly the idea that online commenters are paid to deflect criticism for the government.

Oculus and Anduril founder Palmer Luckey drew attention to the phenomenon on Monday.

In a statement, a YouTube spokesperson told TechCrunch that the auto-deletions were a result of "an error in our enforcement systems" that the company is looking into.

ADVERTISEMENT

"Users can report suspected issues to troubleshoot errors and help us make product improvements," the spokesperson said.

According to YouTube, the situation is an accidental side effect of the platform's comment moderation system, which is designed to filter out hate speech, harassment and spam. The company didn't offer further insight as to how the terms wound up flagged by its automated systems.

With the vast majority of their workforces out of the office, major tech platforms have leaned more heavily on AI moderation methods in recent months, even as they acknowledged that less human oversight would likely lead to more instances of content mistakenly being taken down.