2023-11-17 16:00:40
Not only will content that imitates a singer’s unique voice be removed, but creators will also be required to label AI-generated content starting next year.
YouTube will have two sets of content guidelines for hyperfaking (deepfakes) generated by AI: a very strict set of rules to protect the platform’s music industry partners, and another, more flexible set for everyone else.
YouTube will ask creators in its blog to start labeling AI-generated content “realistic” when they upload videos, and the disclosure requirement is particularly important for topics like elections or ongoing conflicts .
Filters and sanctions that remain to be developed
There is no precise definition yet of what YouTube considers “realistic”, but the video platform will provide more detailed guidance, accompanied by examples, when the disclosure requirement is implemented next year.
YouTube says penalties for not accurately labeling AI-generated content will vary, but might include removal and demonetization. However, it’s unclear how YouTube will know if an unlabeled video was actually generated by AI.
Added a mock presentation label to the description panel. Youtube
Internet users will be able to intervene
From there, things get complicated, enormously. YouTube will allow people to request the removal of videos that “simulate an identifiable person, including their face or voice” using the existing privacy request form.
A content tag generated by YouTube’s artificial intelligence will appear in a video’s description. It will read: “Modified or synthetic content [:] The sound or images have been altered or digitally generated.”
AI-generated labels will appear on videos or in descriptions, depending on the content.
Display label template added to video player. YouTube will ask creators to label AI-generated content “realistic.” Youtube
Absent legal framework on generative AI
All of this risks ending up in court. For example, parody and satire are an important part of the fair use defense in copyright infringement cases, and assessing a person’s publicness is an important part of copyright law. diffamation.
There is no definition of parody and satire for fake videos yet, but YouTube says there will be guidance and examples when the policy is implemented next year.
Easier said than done, YouTube walks a tightrope because there is no established legal framework for copyright in the age of generative AI – there is no law or specific court case that says it’s illegal to train an AI system to sing in Taylor Swift’s voice. But YouTube also depends on the music industry: it needs licenses for all the music that floods the platform daily, in particular to compete with TikTok, which has established itself as the most powerful music discovery tool on the internet. .
Label template added to content generated by YouTube AI Dream Screen. AI tags will appear on videos or in descriptions, depending on the content. Youtube
1700254632
#YouTube #keeping #eye #fake #AIgenerated #musicians
