Authors Guild v. Google: Book Scanning and Snippets as Transformative Fair Use

Authors Guild v. Google, Inc., 804 F.3d 202 (2d Cir. 2015), is the appellate decision that blessed one of the most ambitious copying projects in history — Google’s scanning of more than twenty million books — as fair use. Through its Library Project, Google made full digital copies of books supplied by major research libraries, indexed their entire text so users could search across the corpus, and displayed short “snippets” of text around a searcher’s query terms. Writing for a unanimous panel, Judge Pierre Leval — the jurist whose 1990 article introduced the concept of “transformative” use into fair-use doctrine — held that Google’s search-and-snippet functions serve a transformative purpose that does not supersede the original works and causes no cognizable market harm. The ruling confirmed that copying an entire work, even for profit, can be fair use when it powers a genuinely new and informative function.

At a glance

• Case: Authors Guild v. Google, Inc., 804 F.3d 202 (2d Cir. 2015)
• Court: United States Court of Appeals for the Second Circuit, on appeal from the Southern District of New York
• Decided: October 16, 2015
• Opinion: Judge Pierre N. Leval, for a unanimous panel (Judges Leval, Cabranes, and Parker)
• Subject matter: Fair use of the wholesale digitization of books to create a full-text search index with snippet display
• Holding: Google’s copying to enable full-text search and limited snippet display is a transformative fair use that does not infringe the authors’ copyrights

The Library Project and a decade of litigation

Beginning in 2004, Google partnered with major research libraries to digitize their collections. Libraries delivered physical books; Google scanned each one, applied optical character recognition to render the text machine-readable, and retained a full digital copy. Google returned a digital copy to the contributing library and kept one to power its search index. The public-facing product let anyone search the full text of the entire corpus and, for books still in copyright, view a handful of snippets — roughly an eighth of a page — surrounding the search term, along with information about where to buy or borrow the book.

The Authors Guild and individual authors sued in 2005, alleging massive copyright infringement. After years of litigation, including a rejected class settlement, Judge Denny Chin granted summary judgment to Google on fair use in 2013. The Second Circuit affirmed. Judge Leval structured the opinion around the four statutory factors of 17 U.S.C. § 107, with the transformative-purpose inquiry of the first factor doing the heavy lifting.

Transformative purpose without expressive substitution

The court’s central holding is that Google’s copying was “highly transformative.” Drawing on Campbell v. Acuff-Rose Music and his own foundational scholarship, Judge Leval explained that a use is transformative when it communicates something new and different from the original or serves a new and different purpose, rather than merely superseding the original’s expressive content. Google’s full-text search does exactly that: it does not offer the books to be read but tells users about the books — whether and where particular words and phrases appear, and how frequently — a function that “augments public knowledge” without providing a substitute for reading the works themselves. The court analogized to its prior decision in Authors Guild v. HathiTrust, which had already recognized full-text search as a quintessentially transformative purpose.

Crucially, the court held that copying the entire work was reasonable in relation to that purpose. The third fair-use factor asks whether the amount taken is reasonable relative to the use; because complete digitization is necessary to make full-text search comprehensive and reliable, copying the whole of each book did not weigh against fair use. What mattered was not how much Google copied into its database but how much it revealed to the public — and there, snippet view was tightly limited.

Snippets, market harm, and commercial motive

The court gave careful attention to the snippet-display feature, which posed the greatest risk of expressive substitution. It found that Google’s design contained meaningful safeguards: snippets are small (about one-eighth of a page), the system blacklists one snippet per page and one page in ten so a searcher cannot stitch together a continuous readable portion, and certain reference works and dictionaries where a single snippet could satisfy demand were excluded entirely. Because a user could not reconstruct a book, or even a meaningful part of it, from snippets, the display served the transformative search purpose without becoming a competing substitute for the original.

On the fourth factor — market harm — the court rejected the authors’ theory that snippet view would displace book sales. It reasoned that the brief, discontinuous fragments revealed could not serve as an effective market substitute, and that any lost sales attributable to a searcher finding the specific fact she needed were the kind of harm caused by the dissemination of unprotectable information, not by usurping protected expression. If anything, the court observed, Google Books tends to enhance the sales of books by helping readers identify and locate works they want. Finally, the court dispatched the argument that Google’s commercial character defeated fair use: a profit motive does not preclude fair use, especially where the use is highly transformative, and countless favored uses such as news reporting and scholarship are pursued commercially.

Open questions

Google validated mass digitization for search but expressly reserved harder questions. The court flagged, without deciding, the risk that Google’s stored digital copies could someday be exposed through a security breach, noting that its holding rested on the adequacy of Google’s protections as the record then stood — leaving open how a future data breach might alter the analysis. The opinion also did not resolve how far the transformative-search rationale extends beyond text search into other uses of the underlying digital corpus. Those unanswered questions have become newly urgent as companies invoke Google and HathiTrust to defend copying millions of works to train generative artificial-intelligence models, a use the 2015 court had no occasion to consider.

Implications

• Copying an entire work can be fair use. Where wholesale copying is necessary to enable a transformative function like full-text search, the amount taken does not defeat fair use; what matters is how much protected expression is revealed to the public.
• Function over expression. A tool that provides information about works — search results, word frequencies, locations — serves a different purpose than the works themselves and is strongly transformative.
• Limited display defeats substitution. Technical safeguards that prevent users from reconstructing a meaningful portion of a work keep a display feature from becoming a market substitute.
• Commercial use is not fatal. A profit motive does not preclude fair use when the underlying use is transformative and does not usurp the market for the original.

Frequently asked questions

Why was Google Books ruled a fair use? The Second Circuit found the copying highly transformative because it created a full-text search tool and snippet-display function that provide information about books rather than substituting for them, and it caused no meaningful market harm to authors.

Did displaying snippets change the fair-use result? No. The court held that limited snippet display serves the transformative search purpose, that Google’s design blocks users from reconstructing meaningful portions of a book, and that snippets are not an effective competing substitute for the originals.

Does Google’s commercial motive defeat fair use? No. The court held that a profit motive does not defeat fair use where the use is highly transformative; many transformative uses, including news reporting and criticism, are undertaken for profit.

Authorities and sources

• Authors Guild v. Google, Inc., 804 F.3d 202 (2d Cir. 2015) (decided October 16, 2015). Google Scholar; CourtListener.
• Leval authorship, the transformative-use and snippet analysis, and market-harm holding corroborated by Wikipedia: Authors Guild, Inc. v. Google, Inc..
• Foundational transformative-use standard: Campbell v. Acuff-Rose Music, Inc., 510 U.S. 569 (1994), via Justia; companion decision Authors Guild, Inc. v. HathiTrust, 755 F.3d 87 (2d Cir. 2014).