TikTok Bans AI-Generated Voices in Shopping Livestreams

TikTok has updated its livestreaming rules to prohibit the use of AI-generated voices and non-real-time verbal interactions in shopping broadcasts. The change requires broadcasters to engage directly with viewers using real-time verbal communication or sign language.

Under the updated “Prohibited Content” guidelines, AI-generated voices, audio recordings, or radio-style content are now listed as non-compliant. This restriction applies specifically to TikTok shopping livestreams and appears to address concerns over repetitive or low-quality automated content that may reduce viewer engagement.

The policy contrasts with TikTok’s broader AI initiatives, including AI-powered digital avatars and generative tools. In China, on the domestic Douyin platform, virtual influencers and AI streamers have become a major component of live commerce. TikTok had previously explored similar applications for promotional streams, but the new rules indicate a step back from fully automated audio in shopping contexts.

A potential exception exists for animated figures, as long as they do not cover more than 50% of the screen. However, even in these cases, audio must still meet real-time requirements. The update likely stems from user experience issues, with excessive AI-generated material potentially diminishing the authenticity and appeal of livestream shopping.

This comes right after some other major updates to TikTok, including new AI features. While this might look like a conflict of interest, it seems that the platform is eager to try and crack down on AI being used to mass-generate content through automation, or perhaps to push for more human-led content specifically due to the growing distaste around “ai slop”.