Soulless computer algorithms are already churning out weather bulletins, sports reports, rap lyrics and even passable Chinese poetry.
But it seems machines have now taken another step toward replacing human enterprise by generating their own reviews of serious academic journal papers that are able to impress even experienced academics.
Using automatic text generation software, computer scientists at Italy’s University of Trieste created a series of fake peer reviews of genuine journal papers and asked academics of different levels of seniority to say whether they agreed with their recommendations to accept for publication or not.
In a quarter of cases, academics said they agreed with the fake review’s conclusions, even though they were entirely made up of computer-generated gobbledygook -- or, rather, sentences picked at random from a selection of peer reviews taken from subjects as diverse as brain science, ecology and ornithology.
“Sentences like ‘it would be good if you can also talk about the importance of establishing some good shared benchmarks’ or ‘it would be useful to identify key assumptions in the modeling’ are probably well suited to almost any review,” explained Eric Medvet, an assistant professor in Trieste’s department of engineering and architecture, who conducted the experiment with colleagues at his university’s Machine Learning Lab.
“If, by chance, a generated review combines sentences which are not too specific, but credible, the review itself may appear as written by a real, human reviewer even to the eyes of an experienced reader,” added Medvet, whose paper “Your Paper Has Been Accepted, Rejected, or Whatever: Automatic Generation of Scientific Paper Reviews,” was published in the journal Lecture Notes in Computer Science last month.
Mixing the fake reviews with real reviews was also likely to distort decisions made by academics by making weak papers appear far stronger thanks to a series of glowing reviews, the paper found.
The research team was able to influence the peer review process in one in four cases by throwing fake reviews into the mix, it said.
“This [may be] the situation [faced by a] real conference program chair … who has to take decisions about all the submissions at his or her conference,” Medvet told Times Higher Education.
“He or she could decide [whom to accept] without actually reading all the reviews, or maybe by giving them just a shallow read,” he added.
While computer-generated reviews “cannot possibly deceive any rigorous editorial procedure, [they] could nevertheless find a role in several questionable scenarios and magnify the scale of scholarly frauds,” the paper concludes.
With nearly 1,000 so-called predatory publishers seeking pay-to-publish journal papers, automatically generated reviews may make it easier for bogus papers to gain credibility, he added.
“It is quite easy to spot the fact that [these reviews] are not sound, but not if you do not read them,” said Medvet.