How do we prevent the next wave of research retractions?

Early in my research career, in a laboratory in Japan, I gave my supervisor a set of results I was proud of. He asked me to run the whole experiment again. This wasn’t because he doubted my honesty, but because, as he said, “A result you cannot reproduce is not yet a result.” I was annoyed. But it turned out I had been wrong about one of my measurements. That experience of being asked to be careful is not an insult but a way of learning to respect the work—and it has stayed with me for 25 years now.

I was reminded of the memory because academia in Bangladesh is suffering a difficult period. Recently, an investigation by Prothom Alo revealed that a large number of papers by researchers at our universities have been withdrawn from international journals due to issues such as fabricated data, plagiarism, fake peer review, and even papers brought from commercial “mills.” The number of retractions—at least 325, by the Bangla daily’s estimate, between 2019 and 2026—is worrying, and the first reaction has been to ask who should be punished. That is a fair question. But as both an active researcher and someone involved in quality assurance at my university, I find it more useful to ask this instead: what will trigger the next wave of retractions, and are we ready to prevent it?

Tap here to add The Daily Star as a trusted source

Read more

Dishonest academic practices must stop

In my perspective, that next wave is already visible: the undisclosed use of artificial intelligence (AI) in research. AI is now part of how research is done, from literature review and data analysis to coding and language editing. When used openly and responsibly, it can be a legitimate and useful tool, and there is no point pretending researchers will stop using it. The danger is not AI itself; it is the undisclosed use of AI tools. Journals around the world have already begun retracting papers for undisclosed AI-generated content, false citations that no real source supports, and for results no human author can actually stand behind. A researcher who lets a model produce text, data, or references and then presents it as entirely their own is committing a new version of an old offence, and more and more are being caught doing it. Within a year or two, this may well be a leading cause of retraction in our field, which we should prepare for.

The university I work at has just adopted a policy on the responsible use of AI. It does not ban AI, which would be both impractical and inconsistent with how research now works. Instead, it requires disclosure: a researcher must state which AI tools were used and for what purpose, personally verify everything the tools produce, and remain fully responsible for the final work. It says plainly that AI cannot be used to generate data or be listed as an author. In effect, it draws a clear line between using a tool and surrendering one’s integrity. In effect, it also draws a line before the issue becomes big enough to prompt a retraction notice.

It started when the university brought in a formal plagiarism policy a year ago, enforced through a plagiarism-checking tool. Every research paper, thesis, class assignment, and capstone project is checked for originality, under an institution-wide subscription covering more than 12,000 students and over 500 faculty and staff members. The value of this is more cultural than technical. A student learns in their very first semester that their work will be checked, so that by the time they become a researcher, honesty is not a rule forced from outside but simply the way they have always worked.

Read more

How Bangladesh can make its university research budget work

Structure matters when institutional money is involved. When I apply for funding to publish a paper or attend a conference, the request does not go straight to a finance desk. It first passes my department’s head and dean. It then goes to a school research committee, where internal and external experts look at the research—whether it holds up, not just whether the paperwork is in order. Only after that does it reach the central research office for a final review. Real resources are set aside to support researchers, but not a single taka moves until the work has passed through several independent and expert sets of eyes.

Even after an international journal has said yes, our own reviewers look at the work again. And I have had my own accepted papers sent back to me with tough, uncomfortable, but undeniably useful questions. As an author in a hurry, I have found this friction inconvenient; but I know that it is also essential. It is, I believe, the real reason some institutions come across misconduct so rarely. It is hardly that one university’s researchers are more honest than another’s. Clean records, in the end, are not due to luck; they are the expected result of structure.

I do not write this from outside the problem. I write as a researcher whose own work is checked, and checked again, by his colleagues and his institution every time he publishes—and who, after 25 years, has learned to be grateful for it. The retractions troubling us today could have been prevented, so can the ones coming tomorrow, but only if we build the safeguards now, while there is still time to get ahead of retractions altogether.

Dr Md. Abdur Razzak is professor of electrical and electronic engineering and director of Institutional Quality Assurance Cell (IQAC) at Independent University, Bangladesh.

Views expressed in this article are the author's own. 

Follow The Daily Star Opinion on Facebook for the latest opinions, commentaries, and analyses by experts and professionals. To contribute your article or letter to The Daily Star Opinion, see our guidelines for submission.