A masthead list, not a headline figure
OpenAI's own announcement names the publications rather than a price: current and archived content from The Wall Street Journal, Barron's, MarketWatch, Investor's Business Daily, FN and New York Post; The Times, The Sunday Times and The Sun; and The Australian, news.com.au, The Daily Telegraph, The Courier Mail, The Advertiser and Herald Sun. OpenAI describes the grant as permission to display News Corp masthead content in response to user questions and to enhance its products, and says the arrangement excludes News Corp's non news businesses. [1]
Neither company disclosed terms. The Wall Street Journal, itself a News Corp title, reported the agreement could be worth more than $250 million over five years, in a mix of cash and OpenAI credits for News Corp's own tools, and that figure is what most subsequent coverage repeats. The summary above follows that reporting rather than a disclosed number, because none was published. [2][1]
Beyond training data
Reporting on the deal notes it covers two distinct uses: training OpenAI's models on News Corp's archive, and surfacing News Corp reporting directly inside ChatGPT answers to user queries. Earlier publisher agreements OpenAI had struck, including with the Associated Press, Axel Springer, Prisa Media, Le Monde and the Financial Times, granted training access only. [2]
Atlas interpretation: That distinction is the actual news in a deal with no public number. A training license lets a model absorb style and fact from an archive it will never attribute again. A display license puts a named masthead's reporting in front of a user at the moment they ask a question, which is closer to distribution than to ingestion, and is why News Corp describes the deal as a partnership rather than a data sale. [1]
The road News Corp did not take
The New York Times sued OpenAI and Microsoft in late December 2023, accusing them of using millions of its articles to train models that now compete with it as a source of reliable information. Coverage of the News Corp agreement points to that suit as the alternative path News Corp chose not to take, after its own earlier discussions with OpenAI reportedly did not lead anywhere. [2]
Atlas interpretation: Two large publishers facing the same new user of their journalism picked opposite tools: a lawsuit that asks a court to set the price and the rules, and a negotiated license that sets both privately and keeps the relationship intact. The News Corp deal supplies no evidence about which choice a court would vindicate, only that a five year commercial agreement was the one News Corp judged worth signing instead of litigating. [2][1]
Sources
- A landmark multi-year global partnership with News Corp
OpenAI · May 22, 2024
- OpenAI to start using news content from News Corp. in multiyear deal
Freedom Forum Institute, First Amendment Center at MTSU · May 28, 2024