Thursday, March 20, 2008

Plagiarism detection is fair use

Turnitin is a plagiarism-detection service. Schools using Turnitin require students to submit papers through Turnitin’s system, which compares their papers to Turnitin’s database and, if there’s a suspicious match, generates a report for the relevant teacher. Schools can also choose to allow their students’ papers to be added to the database to improve Turnitin’s ability to detect among-student cheating. Plaintiffs objected to Turnitin’s approach and sued for copyright infringement based on papers they’d been required to submit as a condition of receiving school credit.

The district court ruled against plaintiffs.

The court enforced Turnitin’s clickwrap agreement, which stated that users agreed not to hold it responsible for any damages they suffered from using the site, even though (1) plaintiffs included statements on their submitted work that they didn’t agree to Turnitin’s terms (too late; the clickwrap said it was the sole agreement); (2) plaintiffs were infants when they submitted their papers, and thus able to disaffirm contracts (but they can’t both benefit from the contract and disavow it, and they benefited by submitting their papers—this seems fishy to me, though I’m not an expert in this area, since most if not all contracts signed by minors will have benefited them in the past); and (3) plaintiffs claimed duress (the duress came from third parties, their schools, and Turnitin’s not responsible; anyway, plaintiffs could have gone to private school, or moved, if they objected so strongly to their schools’ plagiarism policies).

Independently, the court found Turnitin to be engaged in a fair use. The purpose of the use was “highly transformative,” as the originals were created for educational and creative purposes, and Turnitin’s use was merely to look for plagiarism. (See Tony Reese’s discussion of transformative purpose versus transformation of the work.) The nature of the work was of lesser importance because Turnitin didn’t use the creative aspects of the work. And Turnitin’s use didn’t diminish the incentive to be creative, but rather protects creativity by deterring plagiarism. Thus, factor two is either neutral or favors fair use. (Not sure what incentives are doing in factor-two analysis, but okay.)

The third factor, the amount of copying, is neutral or favors fair use (!): though Turnitin copies the entire work, Turnitin limits access to copies to cases in which there’s a suspicious match, and the scope of its copying is consistent with its legitimate use, detecting plagiarism. (Usually, the number of people who have access to copies is not a factor in the fair use analysis, because the inquiry is the amount of the copyrighted work copied, not the number of copies created, but especially in new technology cases considering limitation of access certainly makes sense.)

Finally, there was no effect on the market for the work. Turnitin doesn’t make the papers publicly accessible; by preventing plagiarism, Turnitin even preserves the marketability of the papers. Plaintiffs argued that Turnitin destroyed their ability to sell term papers to sites that buy and resell such term papers, but (1) in depositions, they disavowed that market because they saw it as cheating, and (2) “to accept Plaintiffs’ argument would contravene the public benefit underpinnings of the Copyright Act and would run counter to the Copyright Act’s purpose of encouraging creative, original work.” (In other words, that’s a bad market. Term papers ought not to be written for compensation; copyright’s incentive function would be a harm to the kind of creativity that ought to exist in that sphere.) Plaintiffs also hypothesized that using Turnitin might lead to false accusations of plagiarism if they later submitted the same work to a publication that also used Turnitin, but they provided no evidence that this would occur, and Turnitin would allow easy verification that the works were actually their own.

I used Turnitin as an exam question last year; only about half the class deemed Turnitin’s business model to be fair use, which just goes to show, I suppose, that where you stand depends on where you sit. There's something here about the use of copyright claims to express what are really moral objections--many students perceive the system as operating on a presumption of guilt and reinforcing an antagonistic, rather than a cooperative, relationship between teachers and students.

Fortunately for plaintiffs, the court rejected Turnitin's rather aggressive counterclaims. The court didn’t force plaintiffs to indemnify Turnitin for its costs, since that part of its terms wasn’t explicitly part of the clickwrap agreement. And the court rejected Turnitin’s counterclaims for trespass to chattels, violation of the Computer Fraud and Abuse Act, and violation of the Virginia state-law CFAA analogue, which were based on the fact that one plaintiff faked a password to get access, on the ground that these claims all required direct harm to Turnitin’s computer system, which had not occurred even if Turnitin suffered consequential damages.

No comments: