As a seasoned gamer with a decade-long career in creating and uploading video content on YouTube, this news about OpenAI using my videos without permission to train their AI system has left me feeling like a virtual character stuck in an endless loop of exploitation. It’s like being a level boss that the players keep farming for resources without ever giving credit or compensation!
A YouTuber filed a lawsuit against OpenAI, claiming that their videos were utilized to educate its artificial intelligence, marking another legal skirmish in the ongoing fight against technology innovators in this field.
Through the ongoing lawsuit, YouTubers are now part of a broad legal battle concerning the unlawful use of copyrighted content to fuel models like ChatGPT. Those filing lawsuits against AI companies encompass various creative professionals such as artists, authors, news outlets, and music labels.
David Millette filed a lawsuit in a federal court in San Francisco last Friday, basing it on an article published by The New York Times in April about OpenAI’s development of a speech recognition system called Whisper. In late 2021, OpenAI reportedly encountered a supply issue as they had virtually exhausted all available internet text resources. According to Millette’s claim, the company, led by Sam Altman, then created Whisper to transcribe YouTube videos, with the intention of using the transcriptions for training the next iteration of GPT.
It’s alleged that OpenAI breached YouTube’s terms of service by employing their AI model, Whisper, to transcribe over a million hours of video content from YouTube. The terms explicitly forbid users from utilizing the content for standalone applications or gaining access through automated methods such as robots, botnets, or web scrapers. Notably, Greg Brockman – a co-founder and the president of OpenAI who has temporarily stepped back – is credited as one of Whisper’s creators in a research paper.
The complaint notes that datasets used by OpenAI’s Language Models incorporate transcribed videos from YouTube, as these video transcripts represent a vast collection of real-world spoken language data that can be utilized for training and refining the models.
It’s been reported that some Google employees knew OpenAI was using YouTube videos as training data, but chose not to intervene because Alphabet (Google’s parent company) was doing the same for its own AI development. The Times suggests that if Google were to accuse OpenAI of potential copyright infringement against YouTube creators, it could encounter the same backlash, according to sources familiar with the matter.
It’s important to note that Millette doesn’t accuse OpenAI of copyright violation, but instead argues for unjust enrichment and unfair competition related to the use of video transcripts without his permission or payment. He is demanding at least $5 million and an injunction preventing further usage of his content by OpenAI.
On July 30, U.S. District Judge Araceli Martínez-Olguín ruled against a claim made by top authors suing OpenAI, a similar claim that Millette had also advanced. The judge determined that federal law prohibits this claim because it involves copyrighted material. Although some of her reasoning was based on the fact that it shares grounds with a claim for direct copyright infringement (which wasn’t part of the class action representing YouTubers), the main reason for dismissal was due to its connection to copyright subject matter.
https://www.scribd.com/embeds/757014751/content
Read More
Sorry. No data so far.
2024-08-08 00:25