Scholars discover a mysterious wave of scientific papers using odd, “tortured phrases”

Last spring, Guillaume Cabanac—a computer scientist at the University of Toulouse—and two colleagues noticed a weird trend: A surge of published comp-sci academic papers that were using weird phrases in place of well-established ones (previously at Boing Boing).

For example, instead of “artificial intelligence”, the authors would write “counterfeit consciousness”. Instead of “deep neural network” they’d use “profound neural organization,” and they’d say “colossal information” instead of “big data”.

Cabanac and his colleagues did some research and found these weird phrasings in 860 journals, with a cluster of 30 in one single journal, Microprocessors and Microsystems.

They published their findings in a paper in arXiv—and here’s a bigger chart of the curious synonyms they found:

What’s going on? The trio hypothesized that scammy journal contributors are using tools for auto-translating and autogenerating text to crank out articles. Possibly they’re using autotranslation to cover up plagiarism (which is what the folks at Microprocessors and Microsystems, which is doing its own investigation of the awkward articles — and has retracted some of them — suspects). Or maybe they’re using tools like GPT-3 to crank out plausible-sounding articles so they can beef up their CVs.

Nature wrote a story on the findings …

Elisabeth Bik, a research-integrity analyst in California known for her skill in spotting duplicated images in papers, says that the findings of Cabanac’s research are “shocking”. “This is a very new and disturbing type of fabricated paper,” she adds.

Jennifer Byrne, a molecular-oncology researcher at the University of Sydney, Australia, who also works on spotting fabricated papers, says that this is probably the tip of the iceberg because the researchers only looked in depth at one journal from one publisher. “These papers were also found because they were of very poor quality, but there could be more plausible AI-generated papers within the literature that are harder to detect,” she adds.

Obviously if these are scams they should be rooted out. But I gotta say, just on an aesthetic level, I’m kind of charmed by these kooky neologisms! I wish all AI researchers would call their field “counterfeit consciousness”.

Read More

Clive Thompson