内容来源：https://aibusiness.com/generative-ai/the-nyt-sues-perplexity-meta-partners-with-publishers

内容总结：

生成式AI版权之争白热化：《纽约时报》起诉Perplexity，Meta与多家媒体达成合作

在生成式AI技术快速发展的背景下，内容版权归属问题正引发激烈冲突。近日，两大事件凸显了出版商与AI厂商之间的紧张关系：一方面，《纽约时报》对AI搜索公司Perplexity提起版权侵权诉讼；另一方面，社交媒体巨头Meta宣布与CNN、福克斯新闻等多家媒体建立合作伙伴关系。

《纽约时报》在向纽约联邦法院提交的诉状中指控，Perplexity长期在未获授权、未支付报酬的情况下，非法抓取并使用其原创新闻报道内容，甚至存在错误署名行为。该报发言人强调，虽支持AI的伦理发展，但坚决反对其内容被未经许可用于商业产品开发。据悉，双方谈判已持续超过18个月，但未能达成协议。

与此同时，Meta通过与主流媒体合作，为其AI问答功能获取实时新闻内容授权。这一“合作”与“诉讼”并存的局面，反映出当前AI行业在内容使用规则上仍处于模糊地带。自ChatGPT发布以来，从最初的行业恐慌，到如今内容创作者集体通过法律途径寻求补偿，版权博弈已进入新阶段。此前，《纽约时报》起诉OpenAI、Getty Images起诉Stability AI等案件均属同类争议。

法律专家指出，本案关键可能在于Perplexity能否证明其使用行为属于“合理使用”，即其AI模型对原始内容的处理是否具有“转化性”，而非单纯复制。此外，诉讼结果也将为行业树立重要先例。有分析认为，《纽约时报》此举亦可能是一种商业施压策略，旨在影响Perplexity的公众形象与融资前景。

Perplexity方面则回应称，新技术发展历来伴随此类诉讼，但历史证明阻碍创新并非出路。行业观察人士指出，无论是Meta的授权合作模式，还是持续不断的法律纠纷，都表明在AI时代迫切需要建立系统化的内容授权机制，而非依赖个案解决。如何平衡技术创新与版权保护，已成为全球AI治理的核心议题之一。

中文翻译：

由谷歌云赞助
如何选择首个生成式AI应用场景
要着手应用生成式AI，首先应关注能够优化人类信息交互体验的领域。

当前并行发展的态势凸显了AI市场的矛盾：内容出版商正为其用于训练大语言模型的素材寻求补偿。当一家科技巨头与出版商达成合作时，另一家正面临知名出版商的诉讼。

上周五，《纽约时报》起诉生成式AI初创公司Perplexity侵犯版权，指控其非法抓取时报内容，在未经许可或支付报酬的情况下使用原创新闻报道。时报发言人Graham James在官网声明中表示：
"我们虽支持AI的伦理与负责任应用，但坚决反对Perplexity未经授权使用我方内容开发推广产品。对于漠视我们劳动价值的企业，我们将持续追究其责任。"

同日，社交媒体巨头Meta宣布与CNN、福克斯新闻、人物杂志、今日美国等媒体建立合作。Meta称，该协议将使用户通过Meta AI获取新闻相关咨询时，获得更丰富的实时信息。

内容创作者与AI厂商的博弈
无论是《纽约时报》诉Perplexity，还是Meta与出版商的合作，都揭示了生成式AI时代仍悬而未决的核心争议：内容所有权归属与版权界定。

过去三年间，自ChatGPT发布以来，行业焦点已从最初对AI汲取原创内容并重构用途的恐慌，转向众多内容创作者的反击与索赔诉求。其中标志性事件包括《纽约时报》起诉OpenAI未经授权使用其内容训练模型。此后同类诉讼激增，如盖蒂图片社诉Stability AI、作家团体诉Anthropic等。

部分案件已达成和解（如Anthropic同意向作家支付15亿美元赔偿），部分仍在审理，而Perplexity则面临多起诉讼指控。

时报与Perplexity的长期对峙
（图片来源：Michael M. Santiago via Getty Images）
此次诉讼早有预兆：时报曾于2024年10月和2025年7月两次向Perplexity发出停止侵权警告。诉状指出，双方历时18个月谈判未果，Perplexity始终在无授权协议情况下使用时报内容，其搜索引擎还存在错误署名问题，声称发布过本不存在的报道。

芝加哥McCready律师事务所董事总经理Michael McCready分析："《纽约时报》证据充分，此案意义重大。若败诉导致Perplexity继续现行做法，将为其他AI公司开创危险先例。"他认为时报长达18个月的协商努力，正说明其胜诉信心。

但McCready也指出，Perplexity或可主张大语言模型具备"内容转化性"："只需获得一个订阅账号，就能辩称其并非简单转载，而是通过训练将原始材料转化为新思想表达。"本案关键将取决于模型训练方式、产出内容性质，以及是否挤占了时报的市场空间，这些事实问题将由法庭裁决。

Perplexity面临的指控不止于此：上月亚马逊曾威胁起诉其AI购物助手伪装真人用户在网站操作；10月Reddit亦起诉该公司及其他数据抓取方非法窃取内容训练AI模型。

商业策略角力
伊利诺伊大学芝加哥分校数据科学与人工智能战略助理副校长Michael G Bennett认为，时报的诉讼可能是一种商业施压策略，"既试探胜诉可能性，又通过舆论压力影响Perplexity形象"。他补充道，作为需要持续吸引投资者的企业，被历史悠久的权威媒体起诉势必损害Perplexity声誉。

Perplexity传播主管Jesse Dwyer在邮件声明中回应："百年来，从广播、电视、互联网、社交媒体到如今的AI，出版商起诉新技术公司的案例从未间断。所幸这类诉讼从未成功，否则我们可能还在用电报讨论此事。"Bennett指出，基于此立场，Perplexity必须在法庭上强力论证其使用行为属于"合理使用"。

授权协议势在必行
McCready强调，针对Perplexity的诉讼与Meta的合作案例共同表明，内容创作者、出版商与AI厂商亟需建立授权机制，类似近期音乐行业词曲作者与唱片公司达成的协议。"需要建立通用授权框架，既让大语言模型能合法使用内容，又保障创作者获得收益。逐案解决根本不现实。"

您可能还喜欢

英文来源：

Sponsored by Google Cloud
Choosing Your First Generative AI Use Cases
To get started with generative AI, first focus on areas that can improve human experiences with information.
The parallel developments highlight the tension in an AI market in which publishers are seeking compensation for their content being used to train large language models.
As one major tech vendor makes deals with publishers, another is being sued by an influential publisher.
On Friday, The New York Times filed a lawsuit against generative AI startup Perplexity for copyright infringement. The Times accuses Perplexity of illegally crawling its material and using its original journalistic reporting without permission or compensation.
In a statement on the Times' website, spokesperson Graham James said:
"While we believe in the ethical and responsible use and development of AI, we firmly object to Perplexity's unlicensed use of our content to develop and promote their products. We will continue to work to hold companies accountable that refuse to recognize the value of our work."
The same day The Times filed its lawsuit in a New York federal court, social media giant Meta said it is partnering with media outlets including CNN, Fox News, People Inc. and USA Today.
With the agreement, social media users will receive more real-time information when they ask Meta AI questions related to news, according to Meta.
Content Creators Versus AI Vendors
Both the Times' lawsuit against Perplexity and Meta's partnerships with publishers reveal an AI market that is still struggling to determine who owns content and what copyright means in the age of generative AI technology.
Over the last three years, since the release of ChatGPT, the dynamic has shifted from panic about what it means for AI chatbots and systems to be able to draw from original content and repurpose it, to many content creators fighting back and demanding compensation.
One of the most notable battles for compensation was when the New York Times sued OpenAI for using the Times' content to train its models without permission. Since that lawsuit, which is still going on, many others have followed, including Getty Images suing Stability AI and authors suing Anthropic.
Resolutions have been reached in some of the cases. For instance, Anthropic settled its lawsuit and agreed to pay the authors $1.5 billion, others are still awaiting court judgments, and still others, like Perplexity, continue to face a slew of allegations in the form of lawsuits.
The Times Verus Perplexity
Photo credit: Michael M. Santiago via Getty Images
Perplexity's suit against the Times was a long time coming, as the publisher sent the AI search vendor a cease-and-desist notice in October 2024 and another in July 2025.
In its suit, the Times alleges that it engaged in negotiations with Perplexity for more than 18 months, but the vendor continued to use its content without a licensing agreement. The Times also said the Perplexity search engine uses wrongful attribution, claiming it wrote or published something it never did.
"The New York Times has got a very solid case because there's a lot at stake here," said Michael McCready, managing director of McCready Law in Chicago. "If The New York Times loses and Perplexity is allowed to continue whatever they've been doing and accessing the New York Times, that creates a precedent for a lot of other AI companies."
He added that the fact that the Times tried to work with Perplexity over 18 months means the publishing company is confident it will win.
However, Perplexity can also make the argument about the transformative power of large language models (LLMs), McCready continued. "All they need to do is have one subscription," he said. If an AI vendor has one subscription, it can argue that it's not merely republishing the material, but the trained material transforms the original into new expression of ideas.
"A lot of this lawsuit is going to depend on how the LLM was trained, what content it produces and whether it displaces the market for The New York Times," McCready added. "Those are all going to be factual questions that a judge or a jury is going to decide ultimately, and where the evidence is going to land for each of those questions is going to be how the case is going to turn out."
The Times is also not the only company challenging the vendor. Last month, Amazon threatened to sue Perplexity because the AI shopping agent was allegedly shopping for users on Amazon’s website, posing as human users. In October, Reddit filed a lawsuit against the vendor and other so-called data scrapers, accusing them of illegally stealing content and using it to train its AI models.
A Business Tactic
The New York Times lawsuit could also be a business tactic, "to add more pressure to Perplexity to determine whether it can actually win the lawsuit," said Michael G Bennett, associate vice chancellor for data science and artificial intelligence strategy at the University of Illinois Chicago. "And then adding pressure to it in terms of the negative perception that more than a few people will take of Perplexity as a result of this lawsuit."
Perplexity still needs to appeal to investors, so a case brought against it by one of the country's oldest publishers is a blemish, Bennett added.
In an emailed statement, Jesse Dwyer, Perplexity's head of communications wrote, "Publishers have been suing new tech companies for a hundred years, starting with radio, TV, the internet, social media and now AI. Fortunately, it's never worked, or we'd all be talking about this by telegraph."
With this argument, Perplexity will have to make a strong argument in court that its business practices amount to fair use (legal and permitted use of copyrighted material in the pursuit of free expression), Bennett said.
The Need for a Deal
The cases against Perplexity, as well as Meta's move to partner with various publishers, demonstrate a need for a licensing agreement between content creators, publishers, and AI vendors, similar to a recent deal in the music industry between songwriters and record labels, McCready said.
"There needs to be developed some kind of licensing that will allow LLMs access to all of this content, and it will allow the content creators to get some money," he said. "Case-by-case basis is just not practical."
You May Also Like