Tiktok Will Let Users Tone Down The Amount Of Ai Content In Their Feed

TikTok is nan latest societal level to manus users much power complete nan magnitude of AI-generated contented they see. The app is experimenting pinch a caller mounting that will let users to petition to "see less" AI successful their "For You" feed.

The caller toggle will beryllium disposable successful “the coming weeks” successful nan app's "manage topics" conception that lets group personalize nan types of videos much apt to look successful their feed. In a blog post, TikTok says nan power is meant to thief users "dial things down" if they want to spot little AI contented successful their recommendations. The update follows a akin move from Pinterest, which has faced complaints astir AI slop drowning retired authentic images connected nan platform.

TikTok, for illustration astir different platforms, has seen an influx of AI-generated contented successful caller years. The company, which requires users to explanation specified creations, says that location are much than 1.3 cardinal videos pinch these labels. At nan aforesaid time, nan institution notes that its existing methods for detecting AI contented aren't perfect, Currently, TikTok relies connected a widely-used watermarking strategy called Content Credentials that adds metadata to AI-generated content. But these signals tin get harder to observe erstwhile images aliases videos are edited successful different apps aliases copied and reshared to different sites, arsenic often happens pinch viral content.

To reside this, TikTok says that it will besides research pinch an further "invisible watermarking" strategy that could thief it place and explanation AI contented much reliably. "'Invisible watermarks' adhd different furniture of safeguards pinch a robust technological 'watermark' that only we tin read, making it harder for others to remove," nan institution explained.