TikTok Struggles Moderating Low-Resource Language Content

TikTok’s content moderation in Kenya is failing to reliably detect harmful or permissible posts in many African languages, according to Christian Science Monitor reporting from interviews and user accounts between 2023 and 2025. Workers outsourced to Teleperformance say they review roughly 500 videos per nine-hour shift for about $300 monthly, while AI models trained mainly in English miss low-resource languages, prompting mass reporting and wrongful takedowns.
Key Points
- 1Moderators review about 500 videos per shift, often lacking language skills to judge content accurately.
- 2AI moderation performs poorly on low-resource African languages, causing automated misses and reliance on humans.
- 3Practitioners risk wrongful takedowns and censorship concerns; platforms should expand language coverage and review processes.
Scoring Rationale
Addresses industry-wide moderation and language gaps with firsthand reporting; limited by single-source reportage and regional focus.
Sources
Public references used for this report.
Practice with real Ad Tech data
90 SQL & Python problems · 15 industry datasets
250 free problems · No credit card
See all Ad Tech problems

