Vimeo joins TikTok, YouTube and Meta in implementing a way for developers to label AI-generated content. The video hosting service announced Wednesday that developers will now should tell viewers when realistic content is created using AI.
The latest updates to Vimeo's Terms of Service and Community Guidelines be certain that AI-generated, synthetically created, or manipulated videos should not confused with real people, places, or events. This is a notable move for Vimeo, because it has turn out to be increasingly difficult to differentiate between real and faux content created by the advancement of generative AI tools.
Vimeo doesn’t require creators to label content that is clearly unrealistic, reminiscent of animated content, videos with obvious visual effects, or that use AI for minor production assistance. However, videos that feature a star saying or doing something they didn't do in real life, or that show manipulated footage of an actual event or location, must be given an AI content label.
Additionally, the corporate announced that AI content labels will appear on videos that leverage Vimeo's AI tools, reminiscent of a tool that may delete long pauses and speech interruptions.
A transparent label will now appear at the underside of the video indicating that creators have voluntarily disclosed the usage of AI. When uploading or editing a video, creators can check a box for AI-generated content and indicate whether AI was used for audio, visuals, or each.
Vimeo currently leaves it as much as creators to label their AI-generated content, but the corporate is working on an automatic system that may detect AI and label the suitable content.
In an official blog post, CEO Philip Moyer wrote: “Our long-term goal is to develop automated labeling systems that may reliably discover AI-generated content, further increasing transparency and reducing the burden on creators.”
Moyer, who only joined in April, has previously spoken about Vimeo’s stance on AI. In one other blog entryhe told users that Vimeo protects user-generated content from AI corporations by prohibiting the training of generative AI models on videos hosted on the platform. Similarly, YouTube’s Neal Mohan expressly declared that the usage of videos on the platform to coach models – including OpenAI’s Sora – is a violation of the Terms of Service.
