YouTube Toughens AI Rules, Hoping to Calm Musicians' Fears - The Messenger
It's time to break the news.The Messenger's slogan

YouTube Toughens AI Rules, Hoping to Calm Musicians’ Fears

YouTube is throwing musicians a bone with rules that will force AI content to be labeled, plus new tools that will make it easier to remove

YouTube’s new labels will make it easier to tell when content is generated by AIGoogle

Following concerns from music artists, AI is being used to moderate itself on YouTube. YouTube is expanding its artist-focused approach to AI on its platform by making it easy for creators to report deepfakes or other digitally altered content that uses their voices in new contexts. In the coming months, YouTube will implement a privacy request process for creators that will allow them to request deepfake content to be removed from the platform. Approvals depend on the nature of the content and who exactly is being impacted, with more well-known individuals subject to a “higher bar” in YouTube’s words, as well as other factors. 

It doesn’t just have to be the impacted individual who requests AI-generated content removal: YouTube is introducing a feature that lets its music partners in its early AI music experiments ask for removal on behalf of their clients. Other labels and distributors will receive access to the feature in the coming months. 

YouTube also intends to disclose when content is AI-generated so that viewers are aware of it and aren’t misled. Creators get the weight of this responsibility; YouTube will require them to tell their audiences when they’ve used AI tools to create content that looks realistic, like someone saying something they never actually did. AI content on sensitive topics like elections and world conflicts will especially have to have these labels, and creators who do not comply may have their videos removed or face other consequences. 

At the same time, YouTube is using AI to catch videos that violate its community guidelines more quickly and more accurately. Generative AI can adapt to new threats, unlike YouTube’s current systems, which need specialized context to remove harmful content. Right now, AI classifiers detect content that could violate YouTube’s community guidelines on a large scale and more than 20,000 human reviewers confirm or deny the AI’s selections. 

The duality of YouTube’s approach comes as no surprise: In August, YouTube CEO Neal Mohan expressed the potential benefits of AI while also acknowledging its drawbacks. 

"I’m incredibly excited about the opportunity of AI to supercharge creativity around the world, but recognize that YouTube and the promise of AI will only be successful if our partners are successful," Mohan stated at the time.

Businesswith Ben White
Sign up for The Messenger’s free, must-read business newsletter, with exclusive reporting and expert analysis from Chief Wall Street Correspondent Ben White.
 
By signing up, you agree to our privacy policy and terms of use.
Thanks for signing up!
You are now signed up for our Business newsletter.