Random sentences published as papers

Two publishers have had to withdraw more than 120 scientific publications. They appeared to have been written by computers.

The fake papers were discovered by computer analyst Cyril Labbé, writes Nature. The paper appear in more than thirty so-called ‘conference proceedings’. Sixteen ‘articles’ appeared in publications by the German publication company Springer; a hundred were published by the American Institute of Electrical and Electronic Engineers.

Random sentences

The publications were ‘written’ by the software program SCIgen, which pastes random sentences together which have nothing to do with each other. The program was developed in 2005 by researchers from the American university MIT. They wanted to prove how simple it is to submit complete nonsense to conferences. The program can be downloaded for free.

Yes, peer review

Publisher Springer has pointed out that the articles are subjected to peer review, ‘making it more mystifying that the papers were accepted’, according to Nature. The other publisher would not want to say whether the papers have been proofed by fellow scientists. ‘Conference proceedings’ in general are of smaller gravity than journals.

The most of the papers which were created by computers were sent to conferences in China. It is as of yet unclear why the papers were submitted and whether the authors knew about it. HOP