This is not investment advice. The author has no position in any of the stocks mentioned. Wccftech.com has a disclosure and ethics policy.
A sizzling report from 404 Media claims that artificial intelligence based video generation software Jupiter has been using a wide database of scraped videos from YouTube and other sources to train its model. Jupiter is operated by Runway AI, Inc., a $1.5 billion startup with funding from some of the biggest names in the industry, such as Google and NVIDIA. Data gathered by 404 includes a spreadsheet that lists down YouTube channels of some of the world's largest media groups and content creators, along with websites that host pirated content and links that direct to specific videos.
An anonymous source told the publication that the spreadsheet was part of a "company-wide" effort to scrape content from the Internet, and the report is the latest in a series of controversies surrounding firms using creators' data for their model training without paying any fees.
Spreadsheet Used To Feed Data To Crawler That Downloaded Videos Through Proxies Alleges Source
According to the details, the purported spreadsheet contains links to YouTube channels of Netflix, Disney, Sony, Pixar, Vice News and other well known media industry names. It also focuses on content creators such as the popular Apple video blogger Marques Brownlee and lifestyle vlogger Casey Neistat.
Along with the spreadsheet, 404 Media also contacted a source that claims to have worked at Runway. While the identity of this person is confidential, they provide key details for how the firm used the spreadsheet internally to train its video generation AI models.
According to them, Runway used the spreadsheet to feed an open source software that scraped the content from YouTube. Along with the channels, the sheet also linked to links of specific videos as part of a "company-wide effort to find good quality videos to build the model with."
Runway did not respond to 404's requests for comment, and Google redirected the publication to its statements made earlier this year when it had commented in April that OpenAI's Sora video generator using YouTube videos to train its model violates YouTube's rules.
According to the source, Runway also allegedly tasked employees with sifting out videos through keywords that focused on specific kinds of content. Videos were also segregated according to subject matter, with some examples being animated shorts and student films. The scraping wasn't limited only to YouTube; sites that run pirated content were also targeted.
Jupiter is the internal code name for Runway's Gen-3 model and 404's test prompts for Gen-3 generated content similar to the videos that were purportedly scraped from YouTube. The publication adds that the model stopped generating these videos after it contacted Runway for comment.
Runway is one of the more popular AI video generation firms. Its latest funding round was a Series C extension in June last year, which valued the firm at $1.5 billion. In the same month, it also secured a place in TIME Magazine's 100 Most Influential Companies, a list with more than 12 AI firms.
About the author: Ramish is a seasoned technology writer and editor with more than a decade of experience. He specializes in semiconductor fabrication and market analysis. With a background in finance and supply chain management - via his bachelors in Finance and a micromasters in supply chain management from MIT - Ramish combines financial rigor with deep industry insight to deliver accurate and authoritative coverage.
Follow Wccftech on Google to get more of our news coverage in your feeds.
