Dow Jones & Co., parent company of Wall Street Journaland the New York Postboth owned by News Corp., Jeff Bezos-backed generative AI startup Perplexity has sued for copyright infringement, the latest in a series of artificial intelligence-related lawsuits that have met with varying degrees of success.

Perplexity “purports to provide its users with accurate and timely news and information on a platform that, in Perplexity's own words, allows users to skip links to original publishers' websites,” says the federal court lawsuit filed Monday in the Southern District of New York. “Perplexity seeks to achieve this by illegally copying publishers’ copyrighted works on a large scale, diverting customers and vital revenue from these copyright holders. This lawsuit is brought by news publishers seeking redress for Perplexity’s brazen plan to compete for readers while exploiting the valuable content the publishers produce.”

The New York Times Perplexity reportedly recently sent a “cease and desist” letter to prevent access to the publication’s content. It had previously sued OpenAI. Separately, in 2023, Sarah Silverman and a group of high-profile authors sued OpenAI and Meta for copyright infringement because their works and books were illegally downloaded and used to train the company's large language model AI software. A judge dismissed part of the Open AI case alleging unfair business practices. The lawsuit against Meta is ongoing. Other publishers have sued, as have visual artists, even though AI companies are now among the most valuable in the world.

Generative AI systems generate content that mimics natural language in response to a prompt using large language models. LLMs are “trained” on large amounts of content, allowing them to successfully put together sentences and paragraphs so that a reader can understand them.

“News articles, analyzes and editorials are very useful content in this training because of, among other things, their clarity and structure, the editing and quality control they receive, their wide range of topics, their diversity of perspectives, the timeliness of their information, their writing style and their tone,” The suit explained.

These “editions” are machine-generated reproductions of human-generated content, arranged by LLMs and other tools, that summarize and paraphrase the original human-generated content, sometimes even reproducing that content verbatim – what is called Using it in the lawsuit is a substitution and not fair. This is a legal doctrine that allows the use of copyrighted material without the owner's consent under certain conditions, such as when the work is “transformed” and no longer substantially resembles the original.

The lawsuit states: “The illegality of this massive copyright infringement at the submission stage does not depend on whether the individual outputs of Perplexity’s so-called “response engine” are in each case sufficiently similar to plaintiffs’ copyrighted works to constitute identical reproductions of those works . It is sufficient that Perplexity makes large-scale copies of Plaintiffs' works in order to create reproductions and/or derivative content intended as substitutes for Plaintiffs' works.”

Instead of lawsuits, licensing agreements are also possible, although different publishers approach this differently. News Corp recently partnered with ChatGPT developer OpenAI to license its content for specific uses in OpenAI applications.

News Corp. companies point out the obvious. Its articles draw on the dedication, talent, skills and experience of experienced journalists, editors and other professionals. “Undermining the financial incentives to create original content results in less content and/or lower quality content being generated, which also reduces the amount of content available to AI.”

“Perplexity's business is fundamentally different from that of traditional search engines, which also copy large amounts of content into their indexes, but do so solely to provide links to the original websites. In its traditional form, a search engine is a discovery tool that directs searchers to websites such as the pages of The Wall Street Journal or the New York Postwhere users can click to find the information and answers they are looking for. These clicks, in turn, generate revenue for content producers.”

The lawsuit says Perplexity also harms plaintiffs' brands “by falsely attributing to plaintiffs certain content that plaintiffs never wrote or published.” These results are apparently referred to as “hallucinations” in AI circles.

“Perplexity’s hallucinations can falsely attribute facts and analysis to content producers like plaintiffs, sometimes citing a false source and other times simply making up fabricated news stories and attributing them to plaintiffs.”

The lawsuit also alleges trademark infringement.

The publishers of News Corp. said they sent Perplexity a letter in July highlighting the legal issues arising from unauthorized use of copyrighted works and offering to discuss a possible licensing agreement, but “Perplexity hasn't bothered.” made to answer it.”

The lawsuit seeks a jury trial. It asks the court to enjoin any further unauthorized use of plaintiffs' content and requests that such content be removed from Perplexity's search results, databases and archives. Damages for each copyright infringement may be up to $150,000 in some cases; statutory damages up to three times the actual damages; actual damage; and Perplexity's winnings for each violation.