The New York Times has escalated its fight with generative AI companies by targeting Perplexity, the Jeff Bezos-backed AI startup known for producing AI summaries and other output from web content.
According to a document reviewed by The Wall Street Journal, the publisher sent Perplexity a cease and desist letter demanding that it stop accessing and using New York Times content. The letter frames the issue as a licensing and copyright dispute, not simply a disagreement over attribution or search visibility.
A direct challenge to Perplexity
The central demand is clear: The New York Times wants Perplexity to stop using its journalism in AI summaries and other output. The publisher argues that Perplexity has benefited from work it did not license.
In the letter, The New York Times says Perplexity has been “unjustly enriched” by using the publisher’s “expressive, carefully written and researched, and edited journalism without a license.” The letter also says that this conduct violates copyright laws.
That language matters because it puts the dispute in economic terms. The New York Times is not only objecting to the presence of its articles in AI-generated answers. It is arguing that the value of its reporting, writing and editing is being captured by an AI company without permission.
For publishers, that distinction is important. News organizations invest in journalism, then distribute it through products, subscriptions and other channels. If an AI system can turn that work into summaries or answers without a license, the publisher may see the arrangement as a substitute for its own reader relationship.
Part of a wider publisher backlash
The Perplexity letter is not the first conflict between The New York Times and an AI company. The publisher is also suing OpenAI for using content without consent to train ChatGPT.
That separate case shows how the dispute has more than one front. One question is whether publisher content can be used to train AI systems. Another is whether AI products can access and use publisher content to generate summaries and other output for users.
Perplexity has also faced criticism beyond The New York Times. Other publishers have accused the company of unethical web scraping. Those accusations point to a broader concern in media: whether AI tools are respecting the boundaries publishers set around their websites and content.
The issue becomes sharper when paywalled articles are involved. A recent study from Copyleaks, a tool used to check for plagiarism and AI-generated content, found that Perplexity was able to summarize paywalled content from publishers.
That finding adds another layer to the dispute. If an AI summary can deliver the substance of restricted journalism, publishers may argue that the tool weakens the value of the paywall. For readers, the output may look like a convenient answer. For the publisher, it may look like unpaid reuse of protected work.
Why licensing is the core issue
The New York Times letter repeatedly turns on the idea of permission. The publisher’s claim is not that AI summaries are automatically unacceptable in every form. The argument described in the letter is that Perplexity is using New York Times journalism without a license.
That makes licensing the practical center of the debate. A license can define what content may be used, how it may be used, whether summaries are allowed, and how the publisher is compensated. Without one, the publisher is left arguing that the AI company is taking value without authorization.
For generative AI companies, content is a key input. For publishers, content is the product. The conflict is difficult because the same material can be treated as raw material by one side and as finished work by the other.
The source article also notes that Perplexity recently launched an ad-revenue share scheme to give some money back to publishers. That move suggests the company recognizes at least some need to address publisher compensation, even as the New York Times dispute shows that revenue sharing may not settle every concern.
Perplexity says it wants to work with the publisher
Perplexity CEO Aravind Srinivas told The Wall Street Journal that the startup is interested in working with The New York Times. He also said, “We have no interest in being anyone’s antagonist here.”
That response presents Perplexity as open to a commercial or collaborative path. But the cease and desist letter shows that The New York Times is not waiting for informal cooperation. It is pressing the company to stop accessing and using its content in AI summaries and other output.
The standoff captures a central question for AI search and answer products: how can they deliver useful summaries while respecting the work of publishers whose reporting supplies the underlying material?
There is no simple answer in the facts described here. The New York Times is using legal pressure to defend its journalism. Perplexity is pointing to a willingness to work with the publisher and has introduced an ad-revenue share scheme for some publishers. Between those positions sits a larger industry fight over copyright, web scraping, paywalled content and the economics of AI summaries.
For now, the message from The New York Times is direct. If Perplexity wants to use the publisher’s work, the publisher says that use needs permission.