top of page
  • Writer's pictureSarah Ruivivar

OpenAI's Sora Training with YouTube Videos Breaks Rules

Image credits: Eyestetix Studio / Unsplash

The artificial intelligence world is buzzing with controversy.


YouTube's CEO, Neal Mohan, has publicly stated that if OpenAI used YouTube videos to train its AI-powered video creation tool, Sora, it would be a "clear violation" of YouTube's terms of service.


This is Mohan's first public comment on the matter, and he emphasised that he had no direct knowledge of whether OpenAI had indeed used YouTube videos for Sora's development. However, he was clear that if this were the case, it would infringe on the platform's rules.


Creators who upload their hard work onto YouTube have certain expectations, one of which is that the platform's terms of service will be respected. These terms do not permit the downloading of transcripts or video snippets, which would be a direct violation if OpenAI has done so.


 

Want to learn more about AI's impact on the world in general and property in particular? Join us on our next Webinar! Click here to register

 

The debate centres around the materials OpenAI uses to train the AI models underpinning popular content creation products like ChatGPT and DALL-E. Generative AI tools like Sora operate by absorbing a variety of web content, using this data as a foundation to generate new content, including videos, photos, narrative text, and more.


OpenAI's Chief Technology Officer, Mira Murati, previously stated uncertainty over whether Sora was trained using user-generated videos from platforms like YouTube, Facebook, and Instagram. However, reports suggest that OpenAI is considering training its next-generation large language model, GPT-5, on transcriptions of public YouTube videos.


Meanwhile, Mohan assured that Google adheres to YouTube's individual contracts with creators before deciding whether to use videos from the platform in training the company’s own powerful AI model, Gemini. This ensures that the use of the videos as training data aligns with the terms of service or the contract signed by the creator.


As AI technology continues to advance, it's clear that the industry needs to navigate the complex waters of data usage and copyright laws. The outcome of this controversy could set a precedent for future AI developments.



 

Want to learn more about AI's impact on the world in general and property in particular? Join us on our next Webinar! Click here to register

 


Made with TRUST_AI - see the Charter: https://www.modelprop.co.uk/trust-ai

116 views0 comments

Recent Posts

See All

Comments


bottom of page