fēi zhōu非洲yǒu有2000 zhǒng2000种yǔ yán语言，dàn但AInèi róng内容shěn hé审核zhǐ只dǒng懂hěn很shǎo少
Africa has more than 2,000 languages, but the AI content moderation systems on many social media platforms can only handle a few of them, and even fewer African languages are processed reliably.
As a result, many videos and posts made in local languages are not really understood by AI.
It often can only look at the images, listen to the audio, or decide whether to remove content based on user reports.
As a result, some normal content is wrongly deleted, while some harmful information is not detected.
This problem has appeared in both Kenya and Ethiopia.
For creators, journalists, and ordinary users, this is very unfair, because when they post in their native language, they are more likely to run into trouble.
Now, some research teams and companies have started building data and models for languages such as Hausa, Swahili, Yoruba, and Amharic, and the African Union has also called for attention to language diversity.
However, there is still a lot of work to do before AI can truly understand African languages.
If platforms want to serve African users, they cannot keep ignoring these languages.