A new pandemic is spreading, reports Nature—that of “extremely productive authorship”. Some researchers are publishing a paper a week and the number of authors who put out over 60 articles per year has quadrupled in less than a decade. The hotbeds of extreme productivity are unexpected—they’re not the high-technology nations, or those deeply invested in data-based planning, but Thailand and Saudi Arabia. Pakistan figures, too.
Concerns have been raised because researchers who write at superhuman speed are sometimes very, very human. They cheat and they buy papers off academic black markets, and simply write their names in the space provided. This seems to be happening a lot in Thailand, where some researchers have admitted to buying readymade papers off the black market.
However, cheating probably accounts for a very small part of the boom. And publicly-available general artificial intelligence software are not to blame either, because the ramp-up in production volumes was observed from 2016, long before ChatGPT. The causes include increased collaboration in research, but the underlying cause is the ‘publish or perish’ culture that has taken over all the inky trades—academia, publishing and journalism.
If they lived in our time, Margaret Mitchell and Emily Brontë, who had just one classic in them, would have been press-ganged into three-book deals. Harper Lee, who was also a one-book wonder, has been done a disservice by the posthumous publication of Go Set a Watchman.
In journalism, the annual human resource ‘self-assessment’, on which career progress depends, often includes queries about the number of stories published that year, which was never a criterion earlier. It’s a quantitative metric imposed on a qualitative question. The story of Kamala by Ashwini Sarin, which appeared in the Indian Express in 1981 and exposed the rampant trafficking of women in the Hindi heartland, inspired Vijay Tendulkar to write a play. Nicholas Tomalin’s 1966 story ‘The General Goes Zapping Charlie Cong’ seems to have inspired a scene in Apocalypse Now. No one asked these reporters how much they had written in those years. What their stories led to was understood to matter much more.
During the pandemic, online preprint servers revolutionised the manner in which the results of scientific research are shared. It was a necessary innovation when the world’s laboratories were rushing to develop vaccines; it had become imperative to share the findings in real time because time spent in waiting for the formal peer review process would cost lives. But now, the overwhelming volume of work online seems to be detracting from perceptions of its reliability.
Separately, Nature reports that 2023 has been a record year for the retraction of academic papers. Some 10,000 were pulled back by publishers concerned about the perceived integrity of their platforms, who started weeding out questionable papers. Of the number, 8,000 were pulled by Hindawi, a subsidiary of Wiley, which is also terminating imprints. But in the meantime, these dodgy papers have apparently been cited about 35,000 times. It remains to be seen if the record of the academic disciples can be scrubbed, but rooting out the papers themselves is the first step.
Reasons offered for retractions include “concerns that the peer review process has been compromised” and “systematic manipulation of the publication and peer-review process”. Earlier, some researchers did academic-culture-jamming by injecting garbage into articles to highlight poor screening processes. The garbage swept past the gatekeepers. Now, artificial intelligence could be doing the same thing in deadly earnestness, and the ‘hallucinations’ it is prone to could be producing a rash of irrelevant footnotes.
AI ‘hallucinations’ are like machine versions of Stephen Glass, the young New Republic journalist who fabricated stories and, to authenticate them, also created whole websites referring to companies and people cited in those stories. Trumped-up citations are among the red flags signalling that a writer has turned to an AI for help with the details. Glass was exposed by Forbes, and now, academic publishers seem to be doing a similar clean-up job. This cannot have accounted for the origins of publishing overdrive. And there is one real reason for extremely productive authorship: the nature of scientific and technical research has changed. Earlier, papers were typically the work of a single lab at one campus. Now, a single Nobel prize may be awarded to dozens of researchers. Think of the number of people who were involved in the discovery of gravitational waves, an experiment that spanned continents.
Research is more interdisciplinary than ever before for a very good reason nature is unaware that humans have divided the sum of their knowledge into physics, chemistry, the life sciences, math and so on. A phenomenon in a living system may be explained by chemistry, and the forces at work in a chemical reaction may be explained by physics. And math could play a bigger part than the other disciplines in designing the experiment which leads to a discovery.
Ultimately, mass publishing is just another symptom of humanity appraising life, the universe and everything in volumetric terms. From bad habits (“Live life kingsize,” said a tobacco company in more innocent times) to bad food (“Just say supersizes,” said a fast food chain) and the powers of 10 on paychecks (now called ‘packages’). Size matters even when the subject is not actually amenable to quantitative methods. Happiness indices try to reduce the immeasurable to the approved rubric of scientific analysis. Seriously, who can say how happy we are? Even the happiest people may have no idea what makes them tick. It’s best left that way, or they could come to grief.
Pratik Kanjilal
Editor of The India Cable
(@pratik_k)