OpenAI Chief's Ambiguity over Sora Model's Data Origin Amid Copyright Lawsuits
Summary:
OpenAI's CTO, Mira Murati, has shown uncertainty about the origin of data used in the training of the firm's upcoming AI model, Sora. In an interview with The Wall Street Journal, she stated that publicly available and licensed data were used, but couldn't confirm if data from social media platforms were included. She confirmed the use of Shutterstock data for Sora's training. OpenAI has faced multiple legal actions owing to its AI models' training data, with authors and The New York Times alleging that their copyrighted content has been used without consent.
OpenAI's Chief Technology Officer, Mira Murati, has exhibited uncertainty regarding the origin of the data utilized to train the firm's soon-to-be launched video generating AI model, Sora. During a discussion with The Wall Street Journal that took place on March 13, clarity about the data origins for Sora, a model designed to produce videos from written instructions, was not provided by Murati. Murati's response to queries about the foundation of the data was that it was obtained from public resources and licensed data. This was how a firm with an $80 billion valuation was preparing its upcoming model. Joanna Stern, from The Journal, inquired whether the training data for Sora was derived from social media platforms like YouTube, Instagram or Facebook, to which Murati answered she was unsure. Stern then referenced OpenAI's collaboration with Shutterstock, the stock image firm, and wondered if their data had been employed in the training of Sora. To this, Murati held back from elaborating on specific data usage, but reiterated that it originated from public or licensed sources. She later confirmed to the Journal that Shutterstock data was indeed utilized for Sora.
AI models are taught to recognize trends, make forecasts, or comprehend language via large amounts of data known as training data sets. Since joining OpenAI in 2018, Murati has played a central role in multiple successful projects of the firm, including DALL-E 3, an image-generating model, the speech recognition instrument Whisper, and the most recent version of their chatbot, GPT-4. She briefly assumed the role of temporary CEO in November 2023 after Sam Altman was dismissed by OpenAI's board.
Legal proceedings involving the training data of OpenAI's AI models have been initiated against the company. In July 2023, OpenAI was sued by authors Sarah Silverman, Richard Kadrey, and Christopher Golden as they claimed that ChatGPT generates summaries of copyrighted material by these authors. A similar lawsuit was filed in December by The New York Times against Microsoft and OpenAI, alleging copyright infringement as the companies used the newspaper's content in the training of their AI chatbots. A separate class-action lawsuit in California accuses OpenAI of extracting private user data from the internet without obtaining user consent to train ChatGPT.
Published At
3/16/2024 11:47:13 PM
Disclaimer: Algoine does not endorse any content or product on this page. Readers should conduct their own research before taking any actions related to the asset, company, or any information in this article and assume full responsibility for their decisions. This article should not be considered as investment advice. Our news is prepared with AI support.
Do you suspect this content may be misleading, incomplete, or inappropriate in any way, requiring modification or removal?
We appreciate your report.