The Hearst Corp., one of the nation's largest information and media companies, announced Tuesday it has become the latest major publisher to sign a content partnership with artificial intelligence giant OpenAI.
Under the terms of the deal, New York-based Hearst will license its vast archives, containing more than 150 years of material, to the San Francisco-based maker of the ChatGPT generative AI tool. OpenAI will be able to integrate Hearst content into its chatbot and other products.
Hearst publishes more than 40 U.S. newspapers and 20 domestic magazine brands. Included in the deal will be content from the San Francisco Chronicle, Houston Chronicle, Hearst Connecticut Media Group, Esquire, Cosmopolitan, ELLE, Runner's World, Women's Health and other titles.
"As generative AI matures, it's critical that journalism created by professional journalists be at the heart of all AI products," said Hearst Newspapers President Jeff Johnson. "This agreement allows the trustworthy and curated content created by Hearst Newspapers' award-winning journalists to be part of OpenAI's products like ChatGPT -- creating more timely and relevant results."
"Bringing Hearst's trusted content into our products elevates our ability to provide engaging, reliable information to our users," said Brad Lightcap, OpenAI's Chief Operating Officer.
Neither Hearst nor OpenAI disclosed the terms of the deal.
Hearst said in a statement that content appearing in ChatGPT "will feature appropriate citations and direct links, providing transparency and easy access to the original Hearst sources."
Hearst Magazines President Debi Chirichella said the deal "will help us evolve the future of magazine content," adding that "this collaboration ensures that our high-quality writing and expertise, cultural and historical context and attribution and credibility are promoted as OpenAI's products evolve."
Hearst said its other businesses outside of magazines and newspapers -- including Fitch Ratings and a minority stake in ESPN -- are not part of the partnership.
Other major media organizations have struck similar arrangements with OpenAI.
Last year German publisher Axel Springer, owner of Politico, Business Insider, and European properties including Bild and Welt, became the first to sign a deal with the world's most famous AI startup.
That partnership allowed ChatGPT to access Springer's archives, learning from its news articles to answer user queries with more recent, accurate information. That also allowed the bot to attribute its answers to specific articles. The companies did not disclose how much money changed hands.
AI programs such as ChatGPT work by taking in massive amounts of written information from across the internet and, combined with high-powered computing, predicting the most likely next word in a sentence based on a user's prompt.
Those systems work better when trained on high-quality material that has been fact-checked and edited, hence the premium put on the archives of leading news outlets and magazines. Humans also play a significant role in fine tuning what are commonly called large language models.
In May, Wall Street Journal owner News Corp. reportedly entered into a deal with OpenAI worth a quarter billion dollars to similarly access and train on its articles and archives.
Outlets including TIME, the Associated Press, France's Le Monde, and the Financial Times have inked similar contracts with OpenAI. Vox Media and The Atlantic have also come aboard to offer their troves of data.
OpenAI is currently flush with cash, having just closed a $6.6 billion funding round valuing the company at $157 billion, according to the New York Times. The Times reported that the startup is now generating hundreds of millions of dollars per month in revenue, but the intense computations needed to make ChatGPT work mean the company is also burning through cash.
Not every publisher has chosen to work with OpenAI.
The New York Times was reportedly in the midst of negotiating a similar deal last year but the negotiations broke down. Instead, the Times has sued the company in federal court.
The case, which is ongoing, accuses OpenAI of copyright infringement and other violations, saying the startup fed protected New York Times content into its chatbots and profited from the material without just compensation.
Authors including Michael Chabon, Ta-Nehisi Coates and comedian Sarah Silverman have also sued OpenAI in federal court. They say ChatGPT's ability to pull up information about their work is proof the bot was trained on their writing without compensation.
OpenAI has claimed fair use under copyright law. Those cases are also ongoing.
Reach Chase DiFeliciantonio: [email protected]; X: @ChaseDiFelice