In a deal revealing the growing value of news content in the age of artificial intelligence, Amazon will pay The New York Times between $20 million and $25 million annually to use its content for training its models and providing answers in its products.
Article contents:
- Introduction: the price of journalism in the age of AI
- Deal terms: more than just training
- A seismic change in information consumption
- The broader context: the content licensing race
- The Times’ stance: high-quality journalism is worth paying for
- Conclusion: a new model for the relationship between technology and media
Introduction: the price of journalism in the age of AI
Amazon’s deal to license a wide range of The New York Times’ content comes with a substantial financial return for the publisher: $20 million to $25 million annually, according to people familiar with the matter. The financial terms of the multi-year agreement, which were not previously disclosed, provide a window into how publishers and AI companies value news content amid a seismic shift in how consumers search for information online. The annual amount represents approximately 1% of The New York Times’ total 2024 revenue.
Deal terms: more than just training
The two companies announced their deal in May, stating it grants Amazon access to content from The New York Times’ news and cooking products, as well as its sports property, The Athletic. Amazon can use the material to train artificial intelligence models and display summaries and short excerpts of The New York Times content across its products and services, including the Alexa voice assistant.
This was the first AI-related licensing agreement for The New York Times and the first of its kind for Amazon with a publisher.
A seismic change in information consumption
The growing popularity of AI-powered chatbots is turning traditional search traffic and its associated advertising revenue from publisher site visits upside down. Technology companies have used articles to train their artificial intelligence models, drawing from news sites to provide users with real-time answers to queries about current events. This poses an existential threat to publishers that rely on clicks to generate revenue.
The broader context: the content licensing race
Amazon’s deal comes amid a broader race among artificial intelligence companies to secure content deals with publishers. OpenAI has agreements with several publishers, including a deal with News Corp, the parent company of The Wall Street Journal, which could be worth more than $250 million over five years. It also has a three-year agreement with Axel Springer, owner of Business Insider and Politico, valued at between $25 million and at least $30 million.
The structure of these deals varies significantly depending on the products and services offered by the AI companies and how they wish to use the content—for example, to train their models, quickly provide chatbot answers, or integrate news into emerging products.
The Times’ stance: high-quality journalism is worth paying for
Interestingly, The New York Times is currently suing OpenAI and Microsoft for copyright infringement. Its decision to reach an agreement with Amazon demonstrates that it is not opposed to the use of its content by artificial intelligence companies in principle, but insists that it must be fairly compensated.
Meredith Kopit Levien, chief executive of The New York Times, said in a memo to staff when announcing the arrangement in May: “The Amazon deal aligns with our longstanding principle that high-quality journalism is worth paying for.”
Conclusion: a new model for the relationship between technology and media
This deal sets an important benchmark in the evolving relationship between tech giants and the media industry. It recognizes the intrinsic value of trusted news content as a vital component for training and feeding artificial intelligence systems. While legal battles continue, licensing deals like this one may represent a more sustainable path forward, where publishers are compensated for their work while AI companies obtain the high-quality data they need to improve their products.
Frequently asked questions
Why is The New York Times suing OpenAI but making a deal with Amazon?
The difference appears to be in the approach. The New York Times views OpenAI as having used its content without permission or compensation, which it considers copyright infringement. In contrast, Amazon came to the table and offered to pay to license the content, which is the model preferred by The New York Times.
Will artificial intelligence stop reducing traffic to publisher sites?
No, chatbots are likely to continue providing summarized answers, which may reduce clicks. However, these licensing deals provide a new revenue stream for publishers to offset some of these losses.
What does this mean for smaller publishers?
There is concern that major tech companies will only cut deals with the largest publishers, leaving smaller news organizations at a disadvantage. This could lead to further concentration in the media industry.
