Is TurnItIn.com, the anti-plagiarism service that compares submitted student papers with existing papers and other works in its database, a force for good or a thief?
According to the Washington Post, four high school students have sued TurnItIn in US District Court this week. They allege that TurnItIn has violated their intellectual property rights by including their papers in its database, after they specifically instructed the company not to do so when checking their papers.
TurnItIn has an enormous database of papers, both professional and student-generated. Instructors upload the work they collect from students, and TurnItIn compares the submitted work with papers already in its database. If a suspicious match is found, the instructor is notified, and may request to see the original paper to compare it with the new paper.
In the process, every paper an instructor uploads is added to TurnItIn’s database. This is the service’s chief advantage over simply using Google (apart from being more automated); without this massive cache of student papers, the service would be far less useful.
TurnItIn’s revenues are in the tens of millions of dollars per year. However, students assert that they are the ones doing the work that makes TurnItIn profitable. The students who are suing have previously complained about the company’s policy of using submitted papers for comparison against later papers. By law, students retain copyright for writing assignments they complete as part of academic coursework.
A second ethical problem with TurnItIn – one which has long sparked student opposition – is that all papers are checked, not just those suspected by instructors of being plagiarized. This assumption of “guilty until proven innocent” has students on the defensive. According to the earlier WaPo article, students at McLean High School circulated a petition in opposition to the school’s use of TurnItIn. The petition received 1100 signatures.
What is best for students – to expect and teach academic honesty, or to actually enforce it? Are these methods incompatible? Are we placing less trust in students when we check on them?
For its part, TurnItIn’s data indicates that 29% of papers submitted show evidence of significant plagiarism, while 1% of papers are copied in their entirety. The remaining 70% are found to be original, and these original papers are added to the database. It’s unclear whether instructors are uploading all of the papers they receive, or only those they suspect of plagiarism; if many instructors are uploading only papers they suspect of cheating, this would overstate the percentage of students who are cheating – and thus, overstate the need for a service like TurnItIn.
The copyright lawsuit will likely hinge on how TurnItIn stores student papers and what it does with them. Mike Smit, a Canadian Computer Science doctoral candidate who has investigated TurnItIn in some depth, and notes that TurnItIn claims not to store or distribute entire papers, but clearly does both. He also points out that hash algorithms could be used to detect copied sentences or phrases without actually storing student papers.
However, in order to prove cheating, an instructor will generally want to show the offending student the paper from which he or she allegedly copied. TurnItIn does allow for this, though it requires the permission of the instructor – not the student – who submitted the original paper. Mike’s experiments with this feature are very informative.
This lawsuit could mark the end of TurnItIn’s business model, which relies so heavily on stored student papers. However, the anti-plagiarism industry is not necessarily doomed if TurnItIn loses in court; a variety of technical means, such as calculating an MD5 hash for each sentence, could be used to detect plagiarism.