It is a curious quirk of academic practice that what are commonly referred to as plagiarism detection systems do?not actually detect plagiarism.

The primary goal of these systems, rather, is to identify passages of text in the document under scrutiny that are similar to passages contained in the repository of documents used for comparison. Even after many years of development, the on the market do a far-from-perfect job in that regard. They cannot reliably detect paraphrases, for instance, or plagiarism of texts that are in other languages or are electronically inaccessible.

But this is far from these programs¡¯ worst failing. Without careful restrictions on their use, they can become not so much a hindrance as a boon to plagiarists.

Specifically, when universities allow their students to access plagiarism detection tools repeatedly, without any controls, those tools can be used by intentional plagiarists ¨C and, potentially, contract cheating services with which they are in cahoots ¨C to camouflage cheating.

É«ºÐÖ±²¥

Uncontrolled access allows cheats to steadily ¡°increase the originality¡± of their work ¨C in terms of a lower percentage of overall similarity to other texts. They iteratively paraphrase identified cited and uncited passages of text until the overall similarity percentage falls below the desired value (or sometimes a threshold specified by the university). This allows them to ensure that the remaining similarities will not be flagged up when tutors run the submitted text through the software.

Some might argue that this iterative process is the whole point of allowing unrestricted access to the software. Students can hone the originality of their words until they have a legitimate claim to call them their own. After all, few students, especially at undergraduate level, ever have truly original ideas: the essence of their task, in writing an essay, is to assimilate and report the thoughts of others in their own words, before drawing their own conclusions on who is right.

É«ºÐÖ±²¥

Yet a mechanical process of iterative paraphrasing has nothing to do with assimilating other people¡¯s thought in any meaningful, educative sense. It still fulfils the essence of plagiarism: stealing ideas/phrases and passing them off as your own.

In the UK, a lot of universities allow students to run drafts of their work through plagiarism detection software (Turnitin mainly) at least once. But this is not the only example. The Czech Republic has its own system, , which checks undergraduate, postgraduate and doctoral theses against a national repository of theses previously submitted since the system began operating in 2008. It also checks theses against published papers, seminar reports and internet sources, and it is used by more than half of higher education institutions in the country.

But anyone can check their work for similarities to these same sources by using the public system, operated by the same prestigious university. Two types of output are available. Users can receive a free email informing them of the overall percentage of similarity. If they sign up for a free account, they can see all the similarities in detail. And, for a small fee, they can see all the percentages of those similarities.

Such a constellation of systems provides ample opportunity for iterative hiding of plagiarism. I?perceive their creation as a big misstep, which has real potential to degrade students¡¯ moral character and undermine academic integrity. Impunity is guaranteed by the fact that there are no checks on account holders¡¯ identity; cheats can call themselves anything they want.

É«ºÐÖ±²¥

Universities¡¯ motivation in offering their students access to plagiarism detection software is, of course, to educate them about plagiarism. Some students are said not to understand where the line is between legitimate use of sources to inform their argument and illegitimate appropriation of those sources. But surely there is a better way to teach students about academic ethics than by showing them that with a thesaurus and a little patience, that line can just be ignored.

J¨²lius Kravjar is a board member of the European Network for Academic Integrity, an association supporting universities to develop good practices in academic integrity.