Mind Matters Natural and Artificial Intelligence News and Analysis
Two computer keyboards, user typing on one of them
Photo by freestocks.org on Unsplash

AI Is Not a Simple Fix for Plagiarism

The internet speeded up a perennial problem without changing it
arroba Email

As it has changed almost everything else, the internet has certainly changed plagiarism — passing off others’ words or ideas as one’s own. One legal case has been preserved from sixth-century Ireland.

Plagiarism is now much easier both to commit and to detect; a search on “plagiarism checker” will turn up a number of free services. But is it more common today? That’s actually hard to know:

Prior to the internet and especially before word processing, plagiarism was almost as labor-intensive as producing original work. Plagiarism was an unethical tool to shore up perceived writing weaknesses rather than a mere shortcut.

Thanks to the internet, that has changed. While many do plagiarize out of concern over the writing, it’s far more common to see plagiarism caused by laziness, poor time management and generally not caring about the project.

Jonathan Bailey, “5 Ways the Internet Changed Plagiarism” at Plagiarism Today

As Bailey perceptively notes, new citation practices in social media and blogging have also “create a shifting attribution standard, one that mixes community expectations and technical limitations.” Put another way, practices perceived as fair use within an online community may be seen otherwise in more rigorous venues.

St. Columba (521–597) was involved in a copyright battle over a manuscript/
J. R. Skelton (1865–1927) (public domain) 

The internet certainly revolutionized the essays-for-hire industry but the problems there are deeper and more subtle:

In this case, assuming the essay-writing services are actually providing brand-new essays, no one else’s work is being stolen without consent. It is being purchased. Nevertheless, the work is being used without attribution, and the students are claiming credit for work they never did. In short, the students are cheating, not learning.

Richard Gunderman, “Write My Essay, Please!” at The Atlantic (2012)

And then the successful cheats assume responsible positions in society…

Are sophisticated programs to detect plagiarism the long-sought answer? Not really, according to an analyst of fifteen years’ experience with them.

Deborah Weber-Wolff notes that algorithm-based systems produce results that are “are often hard to interpret, difficult to navigate, and sometimes just wrong,” including false positives and false negatives. But a second opinion is rarely sought. If the system produces an “originality score,” evaluators may ignore signals of plagiarism that were not detected by the algorithm, such as an abrupt change in style.

One outcome is that the system’s error is never noted, which makes it look more reliable than it is. She closes an article in Nature with a warning and a plea:

Software cannot determine plagiarism; it can only point to some cases of matching text. The systems can be useful for flagging up problems, but not for discriminating between originality and plagiarism. That decision must be taken by a person. The most important method for finding plagiarism is reading a text and studying the references for inconsistencies. A spot check with an Internet search engine, using three to five words from a paragraph or a particularly nice turn of phrase can uncover copyists. Searching for a reference that looks odd might turn up a source that mangled the reference in the same manner. Only if a text is somehow off, and online searching does not help, should software systems be consulted. In those cases, it’s best to use two or three systems, and to read the reports, not take the numbers at face value.

Debora Weber-Wulff, “Plagiarism detectors are a crutch, and a problem” at Nature

She pleads for scientists to see plagiarism as a social problem and work harder to protect the science literature.

Students can still avoid plagiarism by following accepted rules, some of which may seem counterintuitive to them:

Fortunately for anyone who writes these days, there are ways to avoid plagiarism, including what to do when quoting one’s own earlier work, which is seen as a form of plagiarism: If some of the material you are using for your research paper was used by you in your current class, a previous one, or anywhere else you must cite yourself. Treat the text the same as you would if someone else wrote it. It may sound odd, but using material you have used before is called self-plagiarism, and it is not acceptable.”

6 Ways to Avoid Plagiarism in Research Papers” at WriteCheck beta

If imitation is the sincerest form of flattery, plagiarism amounts to passing ourselves off as experts without tears. It’s not realistic to expect software to detect all of the subtleties.

See also: AI can write novels and screenplays better than the pros! #2 in Top Ten AI hypes of 2018 (Robert J. Marks)

Mind Matters News

Breaking and noteworthy news from the exciting world of natural and artificial intelligence at MindMatters.ai.

AI Is Not a Simple Fix for Plagiarism