DMR News

Advancing Digital Conversations

YouTube CEO Neal Mohan Clarifies Stance on AI Training with YouTube Content

ByHuey Yee Ong

Apr 7, 2024
YouTube CEO Neal Mohan Clarifies Stance on AI Training with YouTube Content

YouTube CEO Neal Mohan Clarifies Stance on AI Training with YouTube Content

YouTube CEO Neal Mohan has stated that utilizing YouTube videos to train OpenAI’s text-to-video generator, Sora, would constitute a violation of the platform’s terms of service.

Mohan, speaking publicly on the matter for the first time, clarified that although he lacks direct knowledge of OpenAI’s practices, such use of YouTube content would breach the platform’s guidelines.

Mohan emphasized the expectations creators hold when uploading content to YouTube, asserting that adherence to the terms of service is paramount. Prohibitions against downloading transcripts or video snippets are clearly outlined in these terms, he added, underscoring their significance in maintaining the integrity of content on the platform.

The training methods employed by OpenAI for its AI models, including ChatGPT and DALL-E, have sparked considerable debate. Sora and similar generative AI tools rely on aggregating web content to enhance their capabilities in generating new content across various mediums, including videos, images, and textual narratives.

In the race to advance artificial intelligence, companies such as OpenAI and Google seek to amass extensive datasets to refine their models and produce higher-quality outcomes. Notably, Google and its subsidiary YouTube, both entities under Alphabet Inc., play pivotal roles in this endeavor.

Despite inquiries, OpenAI, supported by Microsoft Corp, has yet to provide commentary on the matter. OpenAI’s Chief Technology Officer, Mira Murati, previously expressed uncertainty regarding whether Sora underwent training using user-generated content from platforms like YouTube, Facebook, or Instagram.

Recent reports suggest that OpenAI is contemplating training its forthcoming large language model, GPT-5, using transcriptions of publicly available YouTube videos. Mohan outlined Google’s approach to incorporating YouTube content into its AI training, emphasizing the adherence to individual creator contracts. He underscored the importance of ensuring alignment with creators’ contractual agreements before utilizing their content for training Google’s AI models, such as Gemini.

“As a platform, YouTube respects the individual contracts with creators,” Mohan stated. “Though some portion of that YouTube corpus may be used for training models like Gemini, it is done in concert with whatever terms of service or contract that creator has signed beforehand.”

Related News:

Featured Image courtesy of Kevin Mazur/Getty Images for YouTube

Huey Yee Ong

Hello, from one tech geek to another. Not your beloved TechCrunch writer, but a writer with an avid interest in the fast-paced tech scenes and all the latest tech mojo. I bring with me a unique take towards tech with a honed applied psychology perspective to make tech news digestible. In other words, I deliver tech news that is easy to read.

Leave a Reply

Your email address will not be published. Required fields are marked *