An AI tool promises to detect profanity, NSFW content in livestreams

Livestreams can be unpredictable, which is part of the fun, right? Not so much for risk-averse brands considering sponsoring a livestream on Twitch, Kick, or YouTube. A new AI product launching today promises to relieve that advertiser anxiety by yanking ads when livestreams go off the rails.

LiveGuard, from streamer-centric NexTide Media, analyzes advertiser-flagged words, weighing context, language, topics, tone, and emotion. Using those clues, the AI decides whether the content is ok or objectionable and, if the latter, yanks the brand mentions from the stream.

"Brands can fully contextualize their guardrails, choosing to avoid or allow categories such as NSFW discussions, political content, profanity, or other sensitive topics, creating total control over where and how their campaigns appear,” according to a statement from NexTide.

LiveGuard is pitched as novel since it operates live and in-stream, and can make decisions fast.  

Minecraft streamer JeromeASF touted the patent-pending AI tool in a statement: "LiveGuard gives creators and advertisers the confidence to collaborate safely, protecting the live experience without limiting creativity.”

Focused on live media advertising, LiveGuard creator Nextide says its services have been utilized by State Farm, the NFL, and the BET Awards.

Livestreams have a long track record of going awry — there’s a whole Reddit page devoted to live fails. Even Elon Musk was burned by a livestream in August, when he "rage quit" after being continuously cyberbullied.