OpenAI and German media giant Axel Springer have signed a first-of-its-kind deal that would see the AI developer pay to license content from the latter’s news sites.
The deal, announced on Wednesday, is one of the first examples in which a media company has been able to secure payment for use of its content by AI developers. Content from Axel Springer, which owns Politico and Business Insider in the U.S. and several publications in Germany, will be used to train OpenAI’s AI models and will populate answers in ChatGPT, the chatbot that made it famous.
ChatGPT will provide summaries of Axel Springer articles for relevant queries that will include links and proper attribution. “This partnership with Axel Springer will help provide people with new ways to access quality, real-time news content through our AI tools,” OpenAI COO Brad Lightcap said.
Financial terms of the deal were not disclosed.
OpenAI will also compensate Axel Springer for using its content to train the artificial intelligence models that power ChatGPT and its photo-generating sibling DALL-E. The Associated Press has a similar deal with OpenAI, which has access to its archives to train its models. However, it doesn’t encompass any provisions for ChatGPT-generated content, which appears to be a new development with Axel Springer.
Since ChatGPT’s launch, artists, creators, and corporate intellectual property owners have been concerned AI companies were using their work without pay. At the start of the year, a trio of commercial artists sued Stability AI and Midjourney, alleging copyright infringement; in July, a group of authors led by comedian Sarah Silverman sued OpenAI and Meta for allegedly using their books to train their AI models. News media publishers in particular have been concerned over their work being used by AI developers. In October, a trade association representing 2,200 media organizations released a paper alleging their news articles made up the bulk of the data used to train AI models from companies like OpenAI and Google.
Those concerns extend to corporate executives, too. Barry Diller urged courts to redefine “fair use” to exclude AI training from exceptions to copyrighted material. And at a conference in November, Elon Musk, who was a cofounder of OpenAI before leaving in 2018, said that all AI companies were using copyrighted work to train their models and denials they were doing so were a “lie.”
Against that backdrop, Axel Springer took a different approach. This summer, CEO Mathias Dopfner signaled his openness to using AI at his company to pursue his “digital only” vision for the future of media. “We want to be ahead of the curve,” Dopfner told Fortune editor-in-chief Alyson Shontell. “We want to be at the forefront of these innovations.”
He went on to say it was apparent many aspects of journalism were outdated, such as when reporters rewrite articles already published by other outlets, a practice known as “aggregating.” In his Fortune interview Dopfner speculated he would prefer AI take over this sort of work, so he could hire “more investigative reporters, more experts who are reporting on certain topics, more great writers.” The goal for Axel Springer, according to Dopfner, was to “proactively” make these adjustments in the era of AI so they weren’t forced on the company. The newly signed agreement with OpenAI does not include provisions for AI-generated content for Axel Springer’s news sites, a company spokesperson told Fortune.
Dopfner framed the OpenAI deal as an evolution of journalism. “We want to explore the opportunities of AI empowered journalism—to bring quality, societal relevance and the business model of journalism to the next level,” he said.
The deal between OpenAI and Axel Springer is not exclusive, meaning that presumably the news publisher could explore other deals. When asked what they might consist of a spokesperson for Axel Springer replied, “They should, just like this one, value the role of journalistic IP for tech products and the overall relevance of journalism for society.”