TikTok labels AI videos and buries unproven posts ahead of the elections
The platform says its in-app election centre lets it review and flag content that could mislead voters, and that unverified video now carries a label.
TikTok says it is removing harmful misinformation, keeping unsubstantiated content out of its For You feed, and labelling video that it cannot verify, as the country moves through an election period.
The platform’s in-app election centre is doing the reviewing work, according to eNCA (YouTube), which broadcast TikTok’s account of its approach on 22 October 2026. The feed is where many users spend their time, and the company says content that could mislead them and undermine the integrity of the elections is identified and reviewed there.
Labelling is the other half of the approach. TikTok says AI-generated content carries a label telling users what they are looking at, and that video it cannot verify is marked as unverifiable.
What the broadcast does not say is how many pieces of content have been removed, how many labels have been applied, or how quickly reviews are completed. It also does not set out what happens to an account that repeatedly posts material the platform flags.
For voters, the practical effect is a label on the screen and a feed that is meant to carry less unproven material. Whether that changes what South Africans see, share and believe before they vote is a question the platform’s own numbers would have to answer.
The source broadcast is available at https://www.youtube.com/watch?v=hYypiWABaug.
