WITH so much information available at the touch of a keyboard or screen, it’s easier than ever to take the work of others and pass it off as one’s own.

Plagiarism, as the transgression is known, has taken centre stage in a cultural battle involving high-profile figures in US academia, finance and politics. Artificial intelligence (AI) tools are creating new opportunities to reuse work created by others, but they also might provide new ways to detect when someone has done so. 

What constitutes plagiarism?

The word is often used to refer to insufficiently credited written material, but it also applies to images, video and music. In written material, plagiarism includes copying another person’s words verbatim without giving credit, failing to mark direct quotations as such and even replicating someone’s sentence structure.

The Associated Press’ policy prohibiting plagiarism says its reporters “must use original content, language and phrasing”. Some universities recommend using no more than three consecutive words – not counting articles and conjunctions – from another source without giving credit; others say five. Plagiarism standards in academia tend to stress crediting the ideas as well as the words of others. 

Why is plagiarism in the news?

Claudine Gay stepped down as president of Harvard University on Jan 2 after she was accused of plagiarism. Her past scholarship, which focuses on the significance of minority office-holding in US politics, had come under scrutiny by conservative activists after she was criticised for her legalistic testimony to Congress about campus anti-Semitism.

Gay acknowledged “instances in my academic writings where some material duplicated other scholars’ language, without proper attribution”, but she called her forced resignation “merely a single skirmish in a broader war to unravel public faith in pillars of American society”.

GET BT IN YOUR INBOX DAILY

Start and end each day with the latest news stories and analyses delivered straight to your inbox.

Sign UpVIEW ALL

One of her loudest detractors was Bill Ackman, the billionaire hedge fund manager who graduated from Harvard and has been a major donor to the university. Two days after Gay stepped down, Business Insider reported that Ackman’s wife, Neri Oxman, had plagiarised multiple paragraphs in her 2010 doctoral dissertation at the Massachusetts Institute of Technology. She acknowledged the lapse and apologised.

Another of Gay’s chief detractors, US Representative Elise Stefanik, a New York Republican, was accused of plagiarising the work of a colleague in a letter she sent urging Gay and two other college presidents to resign.

What are the consequences of plagiarism?

That depends on setting and severity. Plagiarism isn’t a criminal offence in the US. However, owners of copyrights and trademarks can sue those who use their work without permission.

On college campuses, where plagiarism is a uniquely serious offence, policies vary. Penalties for students caught improperly using other people’s work can range from a reduced grade (a source was cited, but quotation marks were missing) to expulsion (a student submits someone else’s work in their own name).

Outside academia, plagiarism has caused reputational damage.

To name just a few instances: Melania Trump was pilloried by critics when, during her husband Donald’s 2016 presidential campaign, she used first lady Michelle Obama’s words in a speech. The New York Times overhauled its reporting and editing policies in 2003 after a reporter, Jayson Blair, was caught cribbing words from other newspapers. The historian Doris Kearns Goodwin stepped down as a Pulitzer Prize judge in 2002 after she admitted to accidentally using chunks of material from other authors. And Joe Biden’s first run for president was derailed in 1987 after he was caught stealing passages from other politicians and admitted to plagiarising a law review article in law school.

Is plagiarism on the rise?

That’s not clear. In a 2011 survey of college presidents by the Pew Research Center, more than half said plagiarism by their students had increased over the previous decade; 89 per cent of them blamed that on the Internet. But a 2019 study that compared four surveys of students taken five years apart at a single university found a drop in reported plagiarism, followed by a plateau from 2014 to 2019.

Also up for debate is whether the shift to online learning during the Covid-19 pandemic swelled the ranks of plagiarists. CopyLeaks, a company that makes plagiarism detection software, said in a 2022 report that its data suggested plagiarism rose by 126 per cent in the US and 124 per cent in France in the 12 months starting in April 2020 compared with the prior three months. A survey of students in the US and Romania found that while cheating on tests increased with the shift to online learning, plagiarism may not have.

What’s the role of AI in plagiarism?

Generative AI services such as ChatGPT create images, videos, text and other content by ingesting vast quantities of material from across the Internet and using that as a guide to produce something new. Such AI tools are notoriously lax in naming the sources of the material they cite, exposing their users to accusations of plagiarism even if the precise wording of a piece is new.

Though it’s unlikely that authors will spot chunks of their own writing in a piece of AI-generated content, they might find their ideas or their research. There’s also evidence that so-called essay mills – which offer pre-written work to students for a fee – are increasingly using AI to handle their workload, making it easier to produce. Many universities have issued guidance for staff on how to detect when students have used an AI tool such as ChatGPT to do the work for them.

How is plagiarism detected?

Tools such as Turnitin.com, launched in 2000, were designed to counter so-called frat file plagiarism, in which students shared entire essays for reuse. Later programs searched for areas of overlap between the text in question and content found online, an approach that flagged properly cited and improperly cited quotations alike. Since then, tools have added the ability to compare syntax and look for stylistic “fingerprints”. They also can check for content copied from video and audio files. Academic experts stress that decisions on disciplinary action should be made by humans looking at the whole range of circumstances around a given paper, rather than just relying on a plagiarism “score” generated by an algorithm.

Can AI help in detecting plagiarism?

Yes. While plagiarism detection tools that compare text against large databases of published content have existed for many years, AI is making them more reliable. Using natural language processing, deep learning and other tools, these systems can spot plagiarism even when the original language has been paraphrased, reworded or AI-generated. BLOOMBERG