LOS ANGELES (AP) – The Center for Investigative Reporting said Thursday it has sued ChatGPT creator OpenAI and its closest business partner, Microsoft, marking a new front in the news industry’s fight against unauthorized use of content. his in artificial intelligence the platforms.
Non-profit organization, which produces Mother Jones AND discover, said OpenAI used its content without permission and without offering compensation, violating the organization’s journalistic copyright. The lawsuit, filed in a New York federal court, describes OpenAI’s business as “built on the exploitation of copyrighted works” and focuses on how AI-generated article summaries threaten publishers.
“It’s extremely dangerous,” Monika Bauerlein, the nonprofit’s CEO, told The Associated Press. “Our existence relies on users finding our work valuable and choosing to support it.”
Bauerlein said that “when people can no longer develop that relationship with our work, when they no longer encounter Mother Jones or Reveal, then their relationship is with the AI tool.”
That, she said, could “rip the whole foundation of our existence as an independent newsroom out from under us,” while also threatening the future of other news organizations.
OpenAI and Microsoft did not immediately respond to requests for comment on Thursday.
The lawsuit is the latest against OpenAI and Microsoft in Manhattan federal court, where the companies are already fighting a series of other copyright lawsuits. from The New York Times, other media and bestselling authors such as John Grisham, Jodi Picoult and George RR Martin. The companies also face a separate case in San Francisco federal court brought by authors including comedian Sarah Silverman.
Some news organizations have chosen to cooperate rather than fight with OpenAI by signing agreements to receive compensation for sharing news content that can be used to train its AI systems. The latest to do so is Time, which announced Thursday that OpenAI will have access to its “extensive archives from the last 101 years”.
OpenAI and other major AI developers typically do not disclose their data sources, but have argued that making collections of publicly accessible online texts, imaging AND other media THE train their AI systems is protected by the “fair use” doctrine of US copyright law. The CIR lawsuit says a dataset that OpenAI has admitted it used to build an earlier version of its chatbot technology contains thousands of links to the website of Mother Jones, a 48-year-old print magazine that has been published in internet since 1993. But the text used for AI training typically lacked information about a story’s author, title, or copyright notice.
Last summer, more than 4,000 writers signed a letter CEOs of OpenAI and other tech companies accusing them of exploitative practices in building chatbots.
“It’s not a free resource for these AI companies to gobble up and make money from,” Bauerlein told news outlets. “They pay for the office space, they pay for the electricity, they pay the salaries for their workers. Why would the content they ingest be the only thing they don’t (pay for)?”
AP is among the news organizations that have made licensing agreements over the past year with OpenAI; others include The Wall Street Journal and New York Post publisher News Corp., The Atlantic, Axel Springer in Germany and Prisa Media in Spain, French newspaper Le Monde and the London-based Financial Times.
Mother Jones and CIR were founded in the 1970s and merged earlier this year. Both are based in San Francisco, as is OpenAI.
The lawsuit by CIR, also known for its Reveal podcast and radio show, outlines the costs of producing investigative journalism and warns that losing control of copyrighted content will result in less revenue and even fewer journalists to report. important stories in “today’s sparse media landscape. ”
“With less investigative news reported, the cost to democracy will be great,” the lawsuit states.
—
O’Brien reported from Providence, Rhode Island.
—
The Associated Press and OpenAI have a licensing and technology agreement which allows OpenAI access to part of the AP text archives.