OpenAI continues to methodically weave its web in the media. The American giant, creator of ChatGPT, has just signed a licensing agreement with Rupert Murdoch’s media empire. The company specializing in generative artificial intelligence (AI) will be able to exploit all of the articles, including archives, from the group’s titles such as the New York Post, the tabloid The Sun, The Times, and The Sunday Times, as well as The Australian, to train its language models.

This “multi-year” agreement gives the right to OpenAI, and therefore to ChatGPT, “to display content from News Corp titles in response to user questions and to improve its products,” said a joint News press release on Thursday. Corp and OpenAI. According to information from the Wall Street Journal, the amount of the agreement would amount to more than $250 million over five years, part of which will be in the form of credits for the use of OpenAI technology. The robot’s millions of users will also have access to links to the content of New Corp’s titles, whose turnover was around $9.9 billion in 2023.

The agreement between Open AI and News Corp specifies that articles will not be available on ChatGPT immediately after publication, as the titles still hope to generate paid digital subscriptions from their readers…

This is the sixth agreement of this type for OpenAI, which has already managed to attract the powerful British financial daily the Financial Times, the American press agency Associated Press, the owner of the Spanish newspaper El País, Prisa Media, or the German Axel Springer (Bild, Politico…) for financial compensation of 25 million dollars over three years. In France, the newspaper Le Monde chose to go it alone by signing the first licensing agreement in the country with OpenAI a few weeks ago.

Also read: AI: challenged by OpenAI, Google goes on the offensive

From OpenAI to Meta, AI giants are constantly searching for new, relevant libraries of data. Their models only become more efficient if we feed them with even more information. These giants first exploited the content, more or less royalty-free, available on the Internet. But now they need more. Entering into agreements with press publishers also significantly reduces the risk of other disputes arising with the media in the years to come… Above all, they guarantee access to up-to-date information.

Because if the list of agreements between OpenAI and the media grows over the weeks, so does that of legal proceedings… The New York Times was the first to open the ball, against OpenAI and its main supporter, Microsoft, for violation of copyright at the end of 2023. Since then, eight American newspapers, including the Chicago Tribune, followed suit at the end of April. The announcement of the agreement with News Corp also comes the day after a new controversy, after actress Scarlett Johansson accused OpenAI of having voluntarily and without her knowledge copied her voice for its new voice assistant.

More generally, the rise of AI presents numerous challenges for the online media business model. Lately, while Google promises to revolutionize its search engine with IA Overviews, which will provide direct answers to users with article summaries, newspaper sites fear seeing their traffic decrease in the future.