Search Articles
(CN) — Ziff Davis, the parent company of several digital media brands, filed a lawsuit in Delaware federal court against OpenAI for copyright infringement.
In the complaint, the nearly 100-year-old publisher claims the artificial intelligence company has copied its written works verbatim to train the model behind popular chatbot ChatGPT.
By doing so, OpenAI has usurped Ziff Davis’s ability to monetize user interactions through advertising, product sales commissions, and other revenue-producing activities, according to the suit.
The company contends that its many well-known media brands have been tarnished as a result. It argues OpenAI has been falsely attributing them to statements and text Ziff Davis never published, as well as attributing its content to other parties.
“Simultaneously with intentionally and egregiously exploiting the content of Ziff Davis and other commercial web publishers without permission, OpenAI is also actively creating and cultivating a market to license content from publishers,” the company says in its complaint.
It accuses OpenAI of copying, reproducing and storing large numbers of the company’s written works in storage systems for Large Language Model training, also called LLMs. It is a type of artificial intelligence that can understand, generate, and translate human language.
The performance of the learning model relies on its training ability, where developers seek large pools of high-quality, well written textual content copied from massive volumes of preexisting works.
To maximize its new Generative Pre-trained Transformer or GPT model, that can generate human-like text in response to a prompt, OpenAI only scraped web pages that have been curated or filtered by humans.
But the responses generated are often misleading or contain misattributions that deceive users and render them unlikely to engage further with any referenced or linked-to Ziff Davis websites, the company says.
“By collecting this textual content without proper authorization or consent from Ziff Davis and copying the content both onto its servers and into the AI models themselves, OpenAI has built its models and financial success on the piracy of protected content and flagrant copyright infringement,” Ziff Davis says in its complaint, filed Thursday.
According to the suit, OpenAI’s engineers also scraped outbound links from the Reddit social media platform that received a certain score from users as “interesting, educational, or just funny.”
Founded in 2015 and backed by billion-dollar tech investors such as Microsoft and Elon Musk, the California-based AI research organization asserts that its “leading AI models” could not exist without unrestricted access to high quality copyrighted books and articles.
With more than 45 sites globally, Ziff Davis is one of the largest publishers in the U.S., providing various sources of consumer-rich content that includes health and wellness, technology, entertainment, as well as product and gaming reviews.
The company said it holds registrations covering more than 1.3 million works with the U.S. Copyright Office, that cover editorial archives of Everyday Health, What to Expect, MedPage Today, IGN, Eurogamer, Games Industry, Dicebreaker, Mashable, PCMag, CNET, ZDNET, and Lifehacker, and more.
The suit comes amid increasing copyright challenges from various media sectors, including artists and authors, against AI companies. Even Musk himself has sued the company, accusing it of going against the original nonprofit mission by prioritizing profit-driven activities, misleading donors, and misusing contributions.
OpenAI is also facing lawsuits from other news organizations such as The New York Times and New York Daily News, while other publications such as the Washington Post have struck licensing deals that allow ChatGPT to use their content.
Last month, a federal judge in New York rejected a request to toss The New York Times’ copyright suit from OpenAI, which argued it was allowed to use material without permission under a legal doctrine known as “fair use.”
Ziff Davis seeks damages of up to $150,000 per copyrighted work. It is represented by the Delaware law firm Benesch, Friedlander, Coplan & Aronoff.
Follow @MegwiththenewsSubscribe to our free newsletters
Our weekly newsletter Closing Arguments offers the latest about ongoing trials, major litigation and rulings in courthouses around the U.S. and the world, while the monthly Under the Lights dishes the legal dirt from Hollywood, sports, Big Tech and the arts.