Imagine you’re a writer, pouring your heart into a story or a meticulously researched tech review, only to find out someone’s copied it, stripped your name off, and used it to train a machine that’s now spitting out answers to questions worldwide. That’s the crux of the beef Ziff Davis, the media giant behind outlets like IGN, CNET, PCMag, and Everyday Health, has with OpenAI, the folks who brought us ChatGPT. In a lawsuit filed in Delaware federal court on April 24, 2025, Ziff Davis is accusing OpenAI of “intentionally and relentlessly” ripping off its content to train its AI models, and they’re not mincing words about it.

Ziff Davis isn’t a small player. With over 45 media brands, 3,800 employees, and a staggering 292 million monthly user visits, it’s a powerhouse in digital publishing. Its portfolio spans tech hubs like CNET and PCMag, gaming giant IGN, and lifestyle sites like Everyday Health and Lifehacker. According to the 62-page complaint, OpenAI has been helping itself to Ziff Davis’s content—think millions of articles, reviews, and guides—without permission. The lawsuit claims OpenAI “copied, reproduced, and stored” this work to train ChatGPT, creating “exact copies” of articles and even derivatives that power the chatbot’s responses. Ziff Davis says it found hundreds of its articles verbatim in just a small, publicly available sample of OpenAI’s WebText dataset. That’s like finding your diary pages photocopied in someone else’s bestseller.

What’s got Ziff Davis particularly steamed is that OpenAI allegedly ignored their attempts to protect their content. Like many publishers, Ziff Davis uses a robots.txt file—a kind of digital “Do Not Enter” sign that tells web crawlers not to scrape their sites. OpenAI, the lawsuit claims, barreled right past these instructions. Worse, Ziff Davis alleges OpenAI stripped copyright information from the content it hoovered up, essentially erasing the publisher’s ownership. The complaint paints a picture of a deliberate, systematic effort to exploit Ziff Davis’s intellectual property, with the company now seeking “at least hundreds of millions of dollars” in damages, according to sources cited by The New York Times. They’re also demanding that OpenAI stop using their work and destroy any datasets or models built with it—a move that could kneecap ChatGPT’s underpinnings.

OpenAI, unsurprisingly, isn’t taking this lying down. In a statement to Reuters, a spokesperson argued that their AI models “empower innovation, and are trained on publicly available data and grounded in fair use.” The fair use doctrine, a cornerstone of U.S. copyright law, allows limited use of copyrighted material without permission for things like education, research, or transformative works that don’t compete with the original. OpenAI is betting that training its large language models (LLMs) falls under this umbrella, arguing that the way ChatGPT processes and generates responses is transformative, not just a regurgitation of Ziff Davis’s articles.

But here’s where it gets murky. Ziff Davis isn’t claiming ChatGPT is spitting out their articles word-for-word every time (though they say it’s happened). They’re arguing that OpenAI’s entire training process—scraping their content, storing it, and using it to fine-tune the model—violates their copyrights and dilutes their trademarks. The lawsuit points out that OpenAI’s actions could undermine the value of their journalism, especially if AI-generated responses start competing with their sites for clicks and ad revenue. It’s a bit like if a chef used your secret recipe to open a rival restaurant, then claimed it’s fine because they tweaked the spices.

OpenAI’s fair use argument isn’t new. They’ve leaned on it in other lawsuits, with mixed results. For instance, a New York federal judge dismissed a similar case from Raw Story in November 2024, suggesting OpenAI’s use of news articles might not always cross the legal line. But Ziff Davis’s sheer scale—nearly 2 million articles a year—and the specific allegations about ignoring robots.txt and stripping copyright info make this case feel weightier. Plus, the Delaware court might see things differently than New York’s.

Ziff Davis isn’t alone in this fight. They’re joining a growing posse of publishers and creators taking aim at AI companies. The New York Times kicked off the trend in December 2023, suing OpenAI and Microsoft for allegedly using millions of its articles to train ChatGPT, demanding “billions” in damages. Other outlets like The Intercept, Raw Story, AlterNet, and a group of Canadian media companies (including CBC and The Globe and Mail) have filed similar suits. Even authors like Ta-Nehisi Coates and John Grisham have jumped in, accusing OpenAI of pilfering their books for AI training. In India, news agency ANI and outlets tied to billionaires Mukesh Ambani and Gautam Adani are suing OpenAI in a Delhi court, signaling this is a global issue.

The stakes are high. Federal copyright law allows fines of up to $150,000 per willful infringement, and with Ziff Davis claiming millions of articles were misused, the math could get astronomical. Some lawsuits, like the Times’s, have even called for the destruction of ChatGPT’s dataset, a move that could force OpenAI to rebuild its models from scratch using only licensed or public-domain data. That’s not just a logistical nightmare—it could reshape how AI companies operate.

Not every publisher is suing, though. Some, like The Associated Press, The Atlantic, The Financial Times, and The Washington Post, have cut deals with OpenAI, licensing their content for millions. News Corp, which owns The Wall Street Journal, took a similar path. These agreements show there’s a middle ground, but Ziff Davis, like the Times, seems more interested in drawing a line in the sand. As one source told The New York Times, Ziff Davis execs spent months weighing whether to license or litigate before choosing the courtroom.

This lawsuit isn’t just about Ziff Davis and OpenAI—it’s a referendum on how AI interacts with the creative economy. Journalism, already battered by declining ad revenue and newsroom layoffs, relies on intellectual property to survive. If AI companies can scrape content at will, ignoring protections like robots.txt, it could gut publishers’ ability to monetize their work. Ziff Davis’s complaint echoes this fear, arguing that OpenAI’s actions threaten their business model by creating a competitor—ChatGPT—that can summarize or repurpose their content without credit or payment.

On the flip side, AI companies argue that their models are transformative tools that benefit society, from advancing medical research to helping students with homework. OpenAI’s spokesperson said that ChatGPT “enhances human creativity” and solves hard problems. They’ve got a point: LLMs need vast datasets to function, and the internet’s a natural feeding ground. But the question is whether they can keep feasting without permission or if they’ll need to start paying for the buffet.

The courts will have to wrestle with thorny questions. Is scraping a form of reproduction that copyright law protects against? Does training an AI model count as “transformative” under fair use, or is it just a fancy way of copying? And what about robots.txt—does ignoring it breach terms of service or rise to the level of a legal violation? Judges in Delaware, New York, and beyond will likely set precedents that ripple across the AI industry.

Beyond the legalese, there’s an ethical undercurrent here. Ziff Davis’s lawsuit highlights a tension between innovation and fairness. AI has the potential to revolutionize how we work and learn, but should it come at the cost of creators’ rights? Publishers like Ziff Davis invest millions in journalism—reporting, fact-checking, editing—that AI models can’t replicate (at least not yet). If those models are built on the backs of unpaid content, it feels less like progress and more like exploitation.

Then there’s the question of transparency. Ziff Davis’s claim that OpenAI stripped copyright info raises red flags about accountability. If AI companies aren’t upfront about where their data comes from, how can creators trust them? This isn’t just a Ziff Davis problem—it’s a concern for anyone who creates online, from bloggers to indie game devs.

The Ziff Davis-OpenAI showdown is just getting started. The Delaware court will likely see months, if not years, of legal wrangling, especially since OpenAI’s already fighting battles on multiple fronts. A win for Ziff Davis could force AI companies to rethink their data pipelines, prioritizing licenses over scraping. A win for OpenAI might embolden tech giants to keep mining the web, leaning on fair use as a shield. Either way, the outcome could redefine the rules for AI training and content ownership.

For now, Ziff Davis is swinging for the fences, and they’re not alone. As more publishers, authors, and creators join the fray, the pressure’s on AI companies to prove they can innovate without trampling intellectual property. Whether this ends in a landmark ruling or a quiet settlement, one thing’s clear: the days of AI companies treating the internet like a free-for-all are numbered.