Nearly two years after The New York Times dragged OpenAI to court for allegedly stealing millions of its articles, the paper just cut a deal with Amazon to hand over its editorial content for AI training. It's a move that highlights how quickly the media landscape has shifted — and how much money talks when tech giants come knocking.
Key Points:
- Amazon will use NYT, NYT Cooking, and The Athletic content in Alexa and to train its AI.
- The deal excludes Wirecutter, which has an existing affiliate relationship with Amazon.
- The Times is still suing OpenAI and Microsoft for copyright infringement.
- Financial terms of the Amazon deal were not disclosed.
The licensing agreement, announced Thursday, represents a stunning reversal for the Times. Just 15 months ago, the paper walked away from months of negotiations with OpenAI and filed a bombshell lawsuit instead. Now it's handing Amazon the exact type of deal it rejected from ChatGPT's maker.
Here's what makes this so remarkable: OpenAI actually offered the Times a "high-value partnership around real-time display with attribution in ChatGPT" back in December 2023. The companies negotiated for months, with talks continuing until just one week before the Times filed suit. OpenAI proposed letting Times content appear in ChatGPT responses with attribution and links back to the paper — essentially what Amazon is now getting.
But the Times walked away, sued for billions in damages, and demanded the destruction of ChatGPT's entire dataset. The paper claimed OpenAI was trying to "free-ride on The Times's massive investment in its journalism" without proper compensation.
Fast-forward to today: the Times is giving Amazon permission to use its articles, NYT Cooking recipes, and Athletic sports coverage to train AI models and surface content across Amazon's platforms. The arrangement "will bring Times editorial content to a variety of Amazon customer experiences" — which sounds awfully similar to what OpenAI had proposed.
The key difference? Amazon presumably wrote a bigger check— though the companies won't say how much.
OpenAI reportedly offered publishers between $1 million and $5 million annually for content licensing — perhaps not enough to tempt the Times. But News Corp's deal was worth more than $250 million over five years, suggesting the right price can get anyone to the table. The Times clearly found Amazon's offer more appealing than whatever OpenAI put forward.
The media industry has essentially split between those suing and those licensing. Publishers that have cut deals with OpenAI include the Associated Press, Axel Springer, the Atlantic, Condé Nast, Financial Times, Le Monde, News Corp, and Vox Media. The Times led the litigation camp alongside The New York Daily News and Center for Investigative Reporting.
For Amazon, the Times content is part of a desperate sprint to catch up in AI. The company was caught flat-footed when ChatGPT launched in late 2022, scrambling to invest $4 billion in Anthropic and hire talent from AI startups. Amazon's Alexa+ revamp, unveiled in February, costs $19.99 per month but has barely reached users more than six weeks after launch. High-quality journalism could help Amazon make its AI responses more authoritative — assuming anyone actually gets to use them.
The Times' legal strategy now looks more like negotiating theater than principled copyright defense. If AI training without permission is truly copyright infringement — as the paper argues in court — why license the same practice to Amazon? The answer is simple: money talks louder than legal principles.
This isn't the first time the Times has taken contradictory copyright stances. In the 2001 Supreme Court case New York Times Co. v. Tasini, the Times was sued by its own freelance authors for putting their articles in electronic databases without permission or additional payment. The Times argued that requiring author consent would force "massive deletion" of content and harm technological innovation. They prioritized "industry reliance interests" over copyright protections, claiming that enforcing authors' rights would be a "practical impossibility."
The Times lost that case and paid $18 million to settle. But notice the complete reversal: when the Times was using content without permission, they championed broad fair use and warned against stifling innovation. Now that AI companies are using their content without permission, they're demanding strict copyright enforcement and portraying unauthorized use as theft of "deeply human" journalism.
This licensing deal with Amazon positions the Times as both AI industry critic and enabler, a contradiction that might weaken its moral authority in court. It's hard to argue AI training violates "deeply human" journalism when you're simultaneously profiting from that same practice. But in a world where digital advertising revenue has collapsed, pragmatism beats purity — even if it means making peace with the machines you're fighting in federal court.