OpenAI training raises concerns about the use of AI in YouTube videos


OpenAI's text-to-video generator Sora faces criticism over the source of OpenAI training data, the AI ​​company has forayed into the AI ​​space with its latest application Sora.

Its ability to generate video clips from text input has amazed viewers. But new reports have shaken people's trust in the reliability of a massive data set that AlphaGo used to train the innovative AI model.

Image credit: Open AI

OpenAI training data debate

Although OpenAI has not revealed the exact origin of Sora's training data, the problem was compounded when Mira Murati, OpenAI's chief technology officer, admitted in an interview with the Wall Street Journal that she did not know the source. This lack of transparency has been a big problem and is why the tech industry is now very skeptical about any kind of data storage like YouTube.

Neil Mohan, CEO of YouTube, a company owned by Google, warned OpenAI developers. In an interview with Emily Chang on Bloomberg Originals, Mohan revealed that YouTube videos were used in training AI models (like Sora), the use of which violated YouTube's terms of service.

Additionally, he highlighted their commitment to striking deals with YouTube creators and using their content to develop their own AI models, such as the Gemini multimodal AI. However, OpenAI has no official rebuttal to these concerns.

The issue of transparency around usage data and the provenance of AI models is a top priority, especially as AI technologies like Sora become increasingly available to the public. The coming months could see a rise in more topics in this particular discussion as both the dangers and benefits associated with AI-powered video generation are revealed.