Ziff Davis v. OpenAI — S.D.N.Y. Holds Robots.txt Is Not a DMCA Section 1201 Technological Measure

Background

Ziff Davis — the publisher of PCMag, IGN, Mashable, and other digital media properties — sued OpenAI alleging that OpenAI scraped Ziff Davis sites to train its large language models despite the sites’ robots.txt files instructing the OpenAI crawler not to crawl them. In addition to copyright-infringement claims, the complaint added a DMCA § 1201(a) anti-circumvention claim, theorizing that the robots.txt directive is a “technological measure that effectively controls access” to copyrighted works on the Ziff Davis sites and that OpenAI circumvented it by scraping anyway.

OpenAI moved to dismiss the § 1201 claim, arguing that a robots.txt file is, at most, a request — a digital “keep off the grass” sign — with no technological enforcement mechanism, and that ignoring it cannot be circumvention as a matter of statutory construction.

The Court’s Holding

Judge Stein dismissed the § 1201(a) claim. The opinion holds that robots.txt is not a technological measure that “effectively controls access” to a work within the meaning of the statute. To “effectively control access,” the court explained, a measure must, in the ordinary course of its operation, require the application of information, a process, or a treatment to gain access. Robots.txt does none of those things: a crawler can read every byte of the protected site without applying any process to bypass the file. The court rejected the analogy to passwords, encryption, or paywalls.

The court invoked the “keep off the grass” framing repeatedly. A homeowner’s signage tells visitors what is permitted but does nothing to physically block them. A crawler that ignores robots.txt is rude — and may be exposed to other claims, like trespass or breach of contract — but it is not circumventing a technological measure.

Three days later, on December 18, 2025, Judge Stein denied Ziff Davis leave to amend the complaint, treating the § 1201 theory as legally foreclosed rather than curable through repleading.

Key Takeaways

• The § 1201 anti-circumvention statute will not be a workable hook for newspapers and publishers fighting AI scraping. Plaintiffs need to lean on copyright infringement, breach of contract (terms of service), state-law trespass to chattels, hot-news misappropriation, or unfair-competition claims instead.
• The decision sharpens, but does not resolve, the open question of whether ignoring robots.txt has independent legal consequences. The court suggests that contract and trespass theories remain available, while § 1201 does not.
• The opinion is consistent with how robots.txt has been treated in other contexts (e.g., search-engine crawler litigation): a social norm signal, not a technological lock.
• Web publishers concerned about AI scraping should pair robots.txt with actual technical measures — rate limiting, user-agent blocking, IP blocking, authentication walls — if they want a cleaner § 1201 theory.

Why It Matters

The AI-copyright wars have produced a steady stream of theories trying to push beyond direct copyright infringement — the path being litigated in the marquee NYT, Authors Guild, and Bartz cases. Ziff Davis closes one of those side doors. It tells publishers that a robots.txt directive, by itself, will not transform an AI scraper into a federal anti-circumvention defendant.

Practically, this matters because the § 1201 statute carries statutory damages and attorneys’ fees that are independent of the underlying copyright claim, and because it bypasses fair-use defenses entirely. Plaintiffs were hoping to use it as a parallel track to the merits-of-training fight. Judge Stein’s ruling forces them back onto the central battlefield: was OpenAI’s training data ingestion fair use of Ziff Davis copyrighted articles? That fight will continue.