Three left-leaning news websites alleged in a new lawsuit that OpenAI stole their content and used it to train its artificial intelligence model, adding to a string of legal efforts by news media to win compensation from the tech company.
The websites Raw Story and Alternet filed a joint suit in federal court in Manhattan on Wednesday against OpenAI, followed by a similar suit by the Intercept. The two suits allege that OpenAI’s ChatGPT copied their data and violated the Digital Millennium Copyright Act in the process. Other publications, including the New York Times, have also sued OpenAI over copyright claims.
ChatGPT violated the DMCA when it “trained ChatGPT not to acknowledge or respect copyright, not to notify ChatGPT users when the responses they received were protected by journalists’ copyrights, and not to provide attribution when using the works of human journalists,” the joint suit argued.
The reference to the DMCA differentiates Raw Story and Alternet’s suit against OpenAI as it attempts to claim that OpenAI explicitly tried to remove any evidence that humans created the work. The DMCA, which was established in 1998, protects online platforms from being held accountable for hosting copyrighted content if the platform wasn’t aware it was hosting the copyrighted content and will remove the copyrighted content once it is informed. The three companies would have to prove OpenAI was manipulating its content to remove those signifiers identifying the reporting as copyrighted. This is the same tool music labels and content creators use to remove illegal copies of movies or songs from YouTube or TikTok.
Alternet and Raw Story CEO John Byrne said he hired an AI expert to confirm ChatGPT was training on the two news websites’ content, according to the Daily Beast.
“This is a time when nonprofit news organizations are struggling to find sufficient financial support for journalism that holds companies like Microsoft and OpenAI to account. As newsrooms throughout the country are decimated by financial imperatives to cut back, OpenAI reaps the benefits of our content,” Intercept CEO Annie Chabel said in a statement. “We hope this lawsuit will send a strong message to AI developers who chose to ignore our copyrights and free ride on the hard work of our journalists.”
Previous copyright lawsuits have claimed training models like OpenAI violated the DMCA. Andersen v. Stability AI, a class-action suit in which a collective of visual artists tried to hold AI image generators accountable for plagiarism claims, had its claims involving the DMCA dismissed due to the suit failing to prove that images were modified to hide proof of copyright.
The New York Times filed its suit against OpenAI and Microsoft in December. It argued OpenAI had used its content to train ChatGPT’s large language model without permission. This was best illustrated in 100 examples in which the New York Times’s lawyers got ChatGPT to replicate the content of New York Times articles word for word. OpenAI pushed back in court on Tuesday, arguing that the New York Times’s lawyers “hacked” ChatGPT to get it to recreate the examples.
CLICK HERE TO READ MORE FROM THE WASHINGTON EXAMINER
Experts in intellectual property law believe it to be the case with the best chance to prove that OpenAI’s use of internet content breaches copyright law.
OpenAI has attempted to mend its relationship with journalists by creating agreements with news outlets such as Politico and Business Insider to feature their content and to compensate them. The AI company has proposed deals with other publishers, and the deals have averaged between $1 million and $5 million a year in compensation.
