YouTube has announced a change to its policy regarding AI content on the platform. Creators will have to inform viewers about AI content, and the company also wants to put an end to counterfeiting the voices of famous performers.
YouTube announced a change in its approach to AI materials. Content generated with the help of artificial intelligence will be clearly marked and, in some cases, banned and eliminated from the platform. The website actually introduces two sets of rules regarding videos created with AI. The first one applies to all users of the platform. The second one has, among others, protect the music industry. Both will come into force “in the coming months”.
YouTube with new AI rules – materials must be marked
From next year, creators on the platform will be obliged to mark the content they have generated with the help of AI. This information will go into the video description, and the appropriate banner will also appear in the player. The obligation applies only to “credible-looking” content. So far, the company does not provide an exact boundary as to what will be considered realistic and what will not.
The labeling requirement applies to both content created entirely by artificial intelligence and material modified using algorithms. However, the principles also go beyond just generative AI. Clear information must appear on all artificially modified or generated content.
The company specifically mentions cases such as videos showing events that never happened, as well as people who said or did something that did not actually happen.
“This is especially important in relation to sensitive content regarding, for example, elections, ongoing conflicts, public health crises or government officials,” we read in the company’s official statement.
Those who consistently violate the new Community Guidelines may face content removal, suspension of their partnership with the platform, and other measures.
YouTube on AI parodies – is there a ban on AI covers?
Another aspect is the generation of faces and voices of various people, which is becoming easier with the help of AI. This type of content can be used against their will – we have already seen this in the wave of advertisements featuring fake celebrities. Here, the platform plans to expand the existing privacy complaint system. Using it, you can ask the platform to remove content constituting the so-called deepfakes. However, there are exceptions and additional circumstances, such as the special status of satire and parody and how recognizable a person is.
Interestingly, there are virtually no exceptions of this type when it comes to songs and other musical content generated with the help of AI. For some time now, the so-called AI covers. They present artists performing other artists’ music in their own style. For example, the AI-generated voices of Frank Sinatra, Elvis and Freddie Mercury are extremely popular.
According to YouTube, a unique “singer’s way of singing or a rapper’s voice” will soon be protected against AI on the platform. Music labels and distributors of artists’ content are to receive special tools to remove this type of content. However, the classification process must take into account whether the material is a criticism, analysis or reportage about synthetic vocals.