TikTok AI blocks 18 Malaysian media accounts reporting mosque assault
TikTok's automated content moderation system permanently blocked the accounts of 18 Malaysian media organizations reporting on a mosque sexual assault case, unable to distinguish journalistic coverage from prohibited content. Malaysia's Communications Ministry confirmed the AI misjudgment and initiated discussions with TikTok to prevent similar incidents and potentially grant media outlets exemptions from automated moderation.
What happened
TikTok permanently blocked the accounts of 18 Malaysian media organizations reporting on a sexual assault incident at a mosque in Petaling Jaya. The blockedmedia outlets had covered the assault case, in which video footage was recorded by CCTV.
Communications Minister Fahmi stated that TikTok's AI system automatically flagged and blocked the accounts because it could not distinguish between journalistic reporting and regular user-generated content. One outlet, 《当今大马》, had its account briefly suspended but was restored after intervention.
Reported impact
The media organizations reported they had not violated any media ethics codes, noting that the videos did not clearly display the victim's image. Multiple local media outlets were unable to operate their TikTok accounts, affecting their ability to distribute news content on the platform.
Index analysis
The minister acknowledged that TikTok's AI system occasionally makes serious misjudgments and cannot differentiate between news reporting by media institutions and content posted by regular users. TikTok recently strengthened its content moderation AI capabilities, but the system is still producing false positives on legitimate journalistic content.
Governance controls
The Malaysian government requested a meeting with TikTok to clarify how media outlets' accounts operate on the platform and to prevent future incidents. The Communications Ministry proposed granting media institutions exemptions or different status on TikTok, citing their own editorial guidelines and ethical standards that should be recognized separately from general user content policies.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.