Germany’s largest news publisher, Axel Springer, has made a groundbreaking agreement with OpenAI, granting them permission to utilize their articles for the purpose of training ChatGPT.
OpenAI will pay the publisher to use content from titles like Politico and Business Insider for its AI chatbot’s development.
Axel Springer will allow the use of selected news articles to answer users’ questions. The company owns Bild and Welt in its home market.
The articles will be summarized and linked to for more information. The agreement includes content that is paywalled.
The terms of the agreement were not revealed, but according to a source familiar with the matter, it included an upfront payment for previous content and a licensing fee of tens millions of euros each year.
The agreement is the largest licensing agreement ever signed as publishers struggle to protect their intellectual properties against the rapid growth in AI.
Mathias Dopfner is the chief executive officer of Axel Springer. He said, “We are thrilled to have helped shape this global partnership, which is the first of its type, between Axel Springer & OpenAI.
“We are exploring the possibilities of AI-powered journalism, to take quality, societal relevancy and the business model for journalism to the next stage.”
Brad Lightcap is the chief operating officer at OpenAI. He said: “We’re deeply committed to working closely with publishers and creators all over the world, ensuring that they are able to benefit from new revenue models and advanced AI technology.”
The move, however, sparked backlash among some journalists who claimed they were not informed about the deal.
Joe Bambridge is an assistant editor for Politico Europe. He said, “It’s really disappointing that it was announced before we had any discussion with our staff about the implications, how this will affect us or if any revenue will come to us.”
OpenAI already has a deal in place with The Associated Press, which will allow ChatGPT to use their content.
Negotiations are underway with other publishers including Rupert Murdoch’s News Corp., which owns The Times, Sun, and Wall Street Journal.
The media has raised concern that OpenAI, Google and other tech companies have stolen their content in order to train AI without permission or payment.
The Daily Mail is currently preparing for a court battle with Google , over allegations that the tech giant violated copyright by using hundreds of thousands online news articles to train Bard, its chatbot.
The Telegraph revealed this week that Bytedance, the owner of TikTok, has been accused of scraping UK News Sites in order to train its own ChatGPT competitor.
Politicians and regulators have been unable to reach an agreement on a code of conduct regarding AI due to their differing interpretations of the copyright laws.
Many news outlets still view licensing agreements as a sustainable solution.
Owen Meredith is the chief executive officer of the News Media Association. He said that “while much more must be done to protect trusted journalistic content from the potentially harmful effects of AI technology, it’s a welcome acknowledgment of the crucial principle that rights holders have to consent to the use of their material.”
“The ability of the publisher to protect its IP is dependent upon a robust, enforceable copyright regime.” Before scraping and using content, any developer must first obtain permission from a publisher.
Axel Springer announced its landmark deal just days after announcing that it would shut down the Upday news service, and relaunched it as an “AI-driven trend news generator”.