YouTube will now label deepfake videos and AI generated content
YouTube is rolling out a fresh policy that mandates creators to reveal when they employ artificial intelligence (AI) to fabricate realistic content. The platform is launching a novel tool in Creator Studio specifically for this, focusing on content that could potentially be mistaken for genuine people, locations, or events. This initiative is designed to protect users from being deceived by synthetically produced videos as advanced AI tools make it increasingly challenging to differentiate between authentic and counterfeit content.
What are the conditions?
YouTube's updated policy primarily focuses on videos that simulate the appearance of a real person. Creators are now obligated to reveal when they have digitally manipulated content to substitute one person's face with another's or artificially created a person's voice for video narration. Moreover, modifications of actual events/locations, like creating the illusion of a real building being ablaze, must also be revealed. This includes the creation of realistic scenes portraying fictitious major incidents like a tornado approaching an actual town.
Exceptions to disclosure and enforcement actions
The new policy doesn't necessitate creators to disclose clearly fantastical or animated content, like an individual riding a unicorn through an imaginary world. Furthermore, content that utilizes generative AI for production support, like script creation or automatic captioning, is exempt from disclosure. However, YouTube intends to contemplate enforcement actions against creators who persistently opt not to use labels. In certain instances, YouTube itself will append labels if a creator hasn't done so and if the content could potentially mislead viewers.
Rollout of YouTube's AI disclosure labels
YouTube's AI disclosure labels will begin surfacing across all its platforms and formats in the coming weeks. The deployment will commence with the platform's mobile apps before moving on to desktop and television apps. The majority of videos will feature a label in the expanded description. However, for videos dealing with more sensitive subjects like health or news, YouTube will exhibit a more conspicuous label on the video itself to caution viewers about manipulated or synthetic content.
