Plagiarised research passed automated tests, and I detected it – but only because it copied my work

Earlier this year, I published a paper on the ethics of researching military populations.

The core argument was straightforward: the standard rules researchers follow to protect participants – for example, informed consent and voluntary participation – don’t work the same in an institution built on hierarchy and obedience.

A soldier can, as protected by ethics, say no to participating in research. But when their commanding officer has nominated them, the practical reality of saying no is very different from the legal right to do so. My paper explored the tension between ethical rights and lived reality.

A couple of weeks ago I was asked to peer-review a manuscript submitted to a psychology journal on the same topic. It didn’t take long for me to become suspicious. As I read on, I came to realise the safeguards in place to protect research integrity are not keeping pace with the tools that can be used to circumvent them.

From factual errors to reproduced memos

Within the first couple of pages of the manuscript, I recognised my own work.

The manuscript had the same argument as mine, a similar structure and conceptual framework. Most alarmingly though, it contained my reflexive memos, reproduced and paraphrased as though they belonged to someone else.

Reflexive memos are a kind of research diary, in which a researcher documents their personal reflections on their own research: the dilemmas they faced, the decisions they made, the things they noticed that shaped their thinking. Reflexive memos aren’t drawn from the literature; you can’t find them in another paper and reference them. They come from the researcher’s own life.

Mine documented what is was like navigating a 24-month institutional approval process that became an ordeal of lost paperwork, shifting requirements and bureaucratic dead ends. They documented the concept of being “voluntold” – that is, watching defence personnel be put forward for supposedly voluntary training programs, and recognising the unspoken pressure that made refusal practically impossible.

In the memos, I also documented the tension I felt as a clinical psychologist between my professional obligations around confidentiality and the reporting requirements imposed on me as a researcher working within the defence organisation.

These were reproduced as if they had happened to someone else.

The manuscript also got something factually wrong. It reproduced a scenario from my fieldwork on an Australian Defence Force base, describing the force’s values displayed on flags on the main thoroughfare.

It substituted the value of “bravery” instead of the correct value, “courage” – a synonym, yes, but any researcher working in this field would spot that immediately.

A lucky catch

I can’t say with any certainty how the manuscript was produced. Nor am I sure of what happened to the manuscript after I raised my concerns.

What I can say is that the systematic paraphrasing throughout, the basic factual error, and the reference list padded with loosely relevant citations, is consistent with the use of AI.

The editor-in-chief of the journal, after confirming the plagiarism, reached the same conclusion.

The journal ran the manuscript through iThenticate, an industry-standard plagiarism software used by many major academic publishers. It returned an 8% similarity match, below the threshold that would normally prompt editorial concern. The 8% corresponded to my published article. The rest had been paraphrased thoroughly enough to look like original work.

The incentive structures of academic publishing, where the number of papers you publish affects your career progression and your institution’s rankings, create conditions where the temptation to cut corners is real.

The editor-in-chief noted that the humanities and social sciences have so far been relatively unaffected by fake science flooding scientific literature. He told me he hopes the social sciences and humanities will remain relatively spared from this phenomenon, but I suspect this may be changing.

The peer review system worked in this case. But only because the manuscript happened to be sent to the person whose work had been reproduced. That’s luck, not a safeguard.

Plagiarism tools are designed to find matching text. They’re not designed to ask whether the experiences reported in a piece of writing could plausibly belong to the person claiming them. That’s a question only a human reader with a genuine knowledge of the field can answer.

A deeper concern

But there was a deeper concern that really got to me.

When someone plagiarises a literature review, they steal intellectual ideas. When someone plagiarises a methods section, they steal intellectual labour.

But when someone reproduces a reflexive memo and presents it as their own, that isn’t about claiming someone else’s ideas; they’re claiming someone else’s experiences.

They’re essentially saying: “I was there, I felt this, this happened to me”. They were not there, they did not feel it, it did not happen to them.

I’ve spent more than a decade working as a clinical psychologist within defence mental health services. That clinical experience is what drew me to this research in the first place. The ethical tensions I documented in my article came from my work as a researcher, from real moments, my lived experiences.

Reading them reproduced in someone else’s name was a particular kind of violation that I’m not sure our existing language around plagiarism quite captures.