Amazon’s artificial intelligence licensing deal with The New York Times reportedly carries an eight-figure annual price tag.
The tech giant will pay the news company between $20 million and $25 million per year for content to train its AI models, The Wall Street Journal (WSJ) reported Wednesday (July 30), citing unnamed sources.
The previously undisclosed terms of the multiyear deal provide a glimpse into how publishers and AI companies are valuing news content amid a sea change in the way readers seek out information online, the report said. The yearly payments equal almost 1% of the Times’ total revenue for 2024.
Amazon and the Times announced the AI licensing deal in May, saying the agreement would give Amazon access to content from the Times’ news, cooking and sports products to train its AI models.
It was the Times’ first AI-related licensing agreement and the first such arrangement Amazon had inked with a publisher.
Meanwhile, OpenAI has made similar agreements with several publishers, including a deal with WSJ parent News Corp that could be worth more than $250 million over five years. The ChatGPT maker also has a three-year agreement with Axel Springer, owner of Politico and Business Insider, worth between $25 million and $30 million, the WSJ report said.
Google is reportedly working on a project to recruit around 20 national news outlets to license their materials to train its AI.
At the same time, several publishers have sued AI firms for copyright infringement, including The New York Times, which accused OpenAI and Microsoft of improperly using its material in a legal battle that began in 2023.
More recently, AI firm Cohere was accused by a group of media companies in a lawsuit of improperly using at least 4,000 copyrighted works to train its AI model. Cohere called the suit “misguided and frivolous.”
Two other firms, Meta and Anthropic, won court victories in June when judges decided the companies’ use of copyrighted content qualified as fair use. However, several legal experts told PYMNTS that the overall fight isn’t over.
“While the Meta and Anthropic cases were definitely positive for AI developers, they should not be taken as definitive on the issue of fair use in using copyrighted materials to train AI programs,” Thomas McNulty, attorney at the intellectual property firm Lando and Anastasi, said in an interview with PYMNTS this month.
For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.
