Hal Swerissen

Plagiarism is in the news. Spectacularly because the first black woman to be president of Harvard resigned six months into the job after being accused of plagiarism by right wing spear carrier Christopher Rufo as part of the the US culture wars on ‘wokism’ in universities.

Let’s start with the Claudine Gay story. Gay is a black political scientist who became President of Harvard University in middle of 2023. In December last year she walked into trouble at a US Congress House Committee on Education and the Workforce meeting when asked by the republican Chair: “Does calling for the genocide of Jews violate Harvard’s rules of bullying and harassment?” Gay responded that it “depends on the context.” A technically correct, but not well judged response in a game of political cat and mouse fuelled by claims of antisemitism on university campuses. Gay has admitted this was an error.

Seeing blood in the water Rufo followed up by accusing Gay of plagiarism in her doctoral dissertation and several academic papers; claims which had been scuttlebutt for some time. Pressure mounted and Gay resigned amidst a torrent of personal and professional abuse from the antiwokism warriors on the right. A subsequent investigation by Harvard found that Gay had failed to adequately cite sources and had duplicated language or paraphrased the work of others on a several occasions, although not to the extent that she had committed serious academic misconduct.

Gay seems to have been sloppy on some references and paraphrasing and lifting the odd sentence and paragraph from others. Not a good look for a very senior academic, but not a hanging offence for a student or the average academic. More a case of reprimand, revise and rewrite for them.

Unfortunately for Gay, academic leaders are in a different position. Academic institutions are obsessed with plagiarism, so their leaders need to be paragons of virtue and wisdom, particularly in the cauldron of American identity politics where plagiarism is a handy weapon for others to use against you. It’s hard to be the boss and set the plagiarism rules when you are breaking them. Gay is not the first University leader to fall on their pen for plagiarism.

But what is plagiarism and why are academic institutions so obsessed by it?

Playing the Game

Plagiarism is passing off another’s work as your own. This includes writing, talking, presenting images, producing equations and so on. Academic institutions have detailed codes of practice defining plagiarism for staff and students. There are rules for claiming authorship and referencing and attributing the work of others. Most of these relate to written publications. To understand why plagiarism is such a hot button issue for academia you have to understand the centrality of writing as the vehicle for academic identity and prestige.

Writing is fundamental to the academic job description. Academics get paid, admired and promoted for writing. In theory their writing ought to reflect thinking, research, scholarship and originality. In doing so they are meant to be contributing to our understanding of the world through the advancement of science, research and scholarship. By and large they let us know what they’ve done by writing their work up in journals and books.

Whether or not you write and whether anyone reads what you have written have therefore become the central measures of academic performance. But it turns out that whether or not you’ve advanced knowledge and made the world a better place through your writing is hard to judge. It is easier to measure how many things you’ve written than their quality. And there are mountains of the stuff, Estimates suggest two million research papers a year are published and growing.

Higher education Institutions are judged by the publication and citation rates of their staff. The higher the staff publication and citation rate, the greater the institutional rankings and prestige. Global league tables drive student demand, funding and staff choices. Academic publishing has become a game that academics cannot avoid if they want career advancement.

Fortunately for them the digital revolution has made publishing much easier. Online journals have proliferated and print is dying. You are much more likely to find academics in front of their screens than in libraries, which is lucky because ‘massification’ of higher education means there are now far more of them and they probably would not all fit into their institutional libraries.

The result is that a good deal of academic writing has become marginal, incremental and repetitive. Often it is hardly read, even by the peer reviewers responsible for screening it before publication. The average readership of an academic paper is 10 people. Half are not read by anyone apart from the authors and the reviewers.

Citations are no better. Most papers are rarely cited by others. In the humanities over 80 percent of papers are never cited and, of those that are, only 20 percent are actually read. Even the high citation papers are gamed. For example, the easiest way to get a lot of citations is to write a review paper on a controversial issue.

The production of academic dross is aided by a highly variable voluntary peer review system, predatory publishers and promotional incentives for academics to publish. Only the very top journals take quality really seriously. Most of the remaining 30,000 or so are more interested in chasing submissions to keep them going. Critical reflection and scholarship has given way to managerialism and performance indicators.

This is not to say that nothing that is published is any good. Just that the good is swamped by the trivial, marginal and mundane – often boring, badly written and incomprehensible, even to the initiated.

Most successful academics have learned to play the publication game. A systematic review of the accumulated academic marginalia that saves others from having to read it themselves will get you more citations than an original research paper. Apart from writing review articles that everyone will cite, attach yourself to research teams and groups, chase grants to support your research, get postgraduate students and leverage their work, slice and dice your publications, pick your journals and network with and cite the people who write in them. Once something is published, put summaries up on social media, talk about it at conferences and seminars, put out media releases and make sure it’s included in research data bases. Then monitor and promote your performance metrics assiduously.

Generating and claiming credit for ideas through publication is the name of the game. Although it varies from one discipline to another, academics heavily define themselves and their identity through their publication and citation track record. A healthy h index of at leat 20 and preferably 50 is central to survival.

With all this pressure to publish comes the temptation to cut corners and cheat. This is where plagiarism comes in. At the bottom end of the scale is sloppy referencing and paraphrasing other people’s work, which is common. At the top end, there is systematic wholesale appropriation without acknowledgement which is risky and rare. Estimates vary, but a decade ago a study of academic management publications found that about a quarter were in the sloppy camp with 13 percent heading toward the more serious end. The problem is probably worse today.

Some offenders claim inaccurate note taking, pressure of time and being busy. Others seem more likely to have built lifting fragments and half baked paraphrasing into their research and writing practice. Punishments vary from reprimands and requirements to correct and apologise through to sacking.

It’s worth noting, that excessive concern about plagiarism is a particular form of academic pedantry. In many other settings people use other people’s work as a matter of course without attribution. Politicians and business leaders have speech writers. Junior staff produce reports that senior staff present. Some publications like the Economist do not attribute stories so they can claim a collective editorial voice.

This is not to say that concerns about taking credit for other people’s writing do not exist elsewhere. There are plenty of high profile cases where pinching bits of other people’s books, plays, news paper articles and speeches have gotten people into trouble. But academia is unique because of the centrality of writing in determining identity, prestige and success and the detailed institutional rules, incentives and sanctions for regulating writing and attribution to manage the game.

But the Game is Changing

And then came artificial intelligence. Or at least the version based on large language models now capable of generating a good 1000 word summary of just about any academic topic you can think of. True, at the moment AI often ‘hallucinates’ and makes up sources and facts, but it seems only a matter of time before that becomes less of a problem. AI’s are already good at technical tasks like writing code. It is plausible that they will soon automate much of the process for systematic reviews, research analyses and write ups.

And what’s wrong with that? In the past, academics were expected to hand code research data, laboriously calculate their own statistical analyses, maintain card indexes of references, ferret about in libraries to find books and hand search journals. Writing up was done on type writers. Students were expected to learn these techniques. Digital tools now do all these tasks much more efficiently and no one is complaining. AI is just the development of another tool to improve productivity.

Academics who calculate odds ratios, logistic regressions and probability estimates for their research studies using statistical packages like Stata, SPSS or R are expected to reference the tools they have used. They should verify and check the analyses. We don’t complain if they don’t do the calculations themselves. The skill is in knowing when and how to use packages and what the statistics they generate mean. As generative AI tools like Gemini and ChatGPT improve, why shouldn’t they be used for academic research and writing, providing they are appropriately verified and referenced? The skill will be in asking the right questions and verifying and evaluating the responses.

Of course this will change the nature of academic research, scholarship and publishing. New rules, regulations and incentives will be required. AI is already having a major impact on teaching and assessment and many higher education institutions are moving out of the horse and buggy era of lecturing, tutorials and essays. Moving to contemporary teaching and assessment models in universities is long over due.

For academic publication, it seems inevitable that AI will become a standard tool in the preparation and writing of academic papers. For the moment it is mainly useful as an adjunct – a research assistant to answer and check questions and produce drafts. But it is likely AI will do progressively more of the everyday research and writing tasks and humans will spend more time on framing the questions and analyses and verifying and editing the AI generated responses.

Similarly, reviewers and publishers are almost certain to use AI tools to review, verify and critique submissions they receive. It is already possible to feed papers into generative AIs and get them to provide a critical review. Currently, the products are patchy. They provide pretty good summaries of the research paper, with more variable insights into the methodology, discussion and conclusions. But this is likely to improve fast.

Getting hot and bothered about AIs taking over human tasks, including generating original questions, critical analyses and conclusions misses the point. The development of new tools and techniques has been going on for a long time. Research, science and scholarship is meant to improve knowledge and if that involves using better tools, so what. But this does mean that what counts as research, scholarship and original work will have to change and the idea of plagiarism will have to be adapted. Not every word has to be human produced.

The impact of AI is still emerging, but it looks as if new incentives and rules for academic publishing to match the emergence of AI will have to be developed. Writing may become less central to academic prestige and identity and plagiarism may have to be redefined.