Science publisher pulls fake work


Sat, Mar 01, 2014 - Page 6

Publisher of science journals Springer on Thursday said it would scrap 16 papers from its archives after they were revealed to be computer-generated gibberish.

The fake papers had been submitted to conferences on computer science and engineering whose proceedings were published in specialized, subscription-only publications, Springer said.

“We are in the process of taking down the papers as quickly as possible,” the German-based publisher said in a statement. “This means that they will be removed, not retracted, since they are all nonsense.”

“We are looking into our procedures to find the weakness that could allow something like this to happen, and we will adapt our processes to ensure that it does not happen again,” Springer added.

The embarrassing lapse was exposed by French computer scientist Cyril Labbe of the Joseph Fourier University in Grenoble.

He also spotted more than 100 other “nonsense” papers unwittingly published by the New York-based Institute of Electrical and Electronic Engineers (IEEE), the journal Nature reported.


In a statement, the institute said it had been advised “there might have been some conference papers published in our IEEE Xplore digital library that did not meet our quality standards.”

“We took immediate action to remove those papers, and also refined our processes to prevent papers not meeting our standards from being published in the future,” it said.

Labbe, 41, has been exploring how to detect fake papers written with a program called SCIgen.

At the press of a button, the program cranks out impressive-looking “studies” stuffed with randomly-selected computer and engineering terms.

Here is an example: “Constant-time technology and access points have garnered great interest from both futurists and physicists in the last several years. After years of extensive research into superpages, we confirm the appropriate unification of 128-bit architectures and checksums.”

This “paper” comes complete with fake graphs and citations — essential features in scientific publishing — that in SCIgen’s case includes recent references to famous scientists who died decades or centuries ago.


The program was devised in 2005 by researchers at the Massachusetts Institute of Technology. They used it to concoct meaningless papers that were accepted by conferences. The researchers later revealed the hoax to expose flaws in safeguards.

Labbe said he spotted the frauds by searching for telltale SCIgen vocabulary.

The fake papers detected by Labbe were submitted to conferences between 2008 and last year. They were uncovered through research he published in 2012 in Scientometrics — by coincidence, also a Springer journal.

In some cases, he said, a paper’s introduction or conclusion were rewritten by a human to appear more authentic at first glance — a veneer presumably aimed at fooling superficial scrutiny.

Labbe said the fraud struck at the credibility of peer-reviewed systems in which scientific claims are meant to be assessed by independent experts for soundness.

“There are several possible explanations” for the fakes, he said.

“One is that people are just testing the system, but if that’s the case, they should reveal who they are and they haven’t done so,” Labbe said. “Another is that the papers are a deliberate fraud to make money.”

Springer said scientific publishing, like other fields, “is not immune to fraud and mistakes.”

“The peer-review system is the best system we have so far and this incident will lead to additional measures on the part of Springer to strengthen it,” the firm said.