Tortured phrases: A dubious writing style emerging in science. Evidence of critical issues affecting established journals

Abstract: Probabilistic text generators have been used to produce fake scientific papers for more than a decade. Such nonsensical papers are easily detected by both human and machine. Now more complex AI-powered generation techniques produce texts indistinguishable from that of humans and the generation of scientific texts from a few keywords has been documented. Our study introduces the concept of tortured phrases: unexpected weird phrases in lieu of established ones, such as 'counterfeit consciousness' instead of 'artificial intelligence.' We combed the literature for tortured phrases and study one reputable journal where these concentrated en masse. Hypothesising the use of advanced LLMs we ran a detector on the abstracts of recent articles of this journal and on several control sets. The pairwise comparisons reveal a concentration of abstracts flagged as 'synthetic' in the journal. We also highlight irregularities in its operation, such as abrupt changes in editorial timelines. We substantiate our call for investigation by analysing several individual dubious articles, stressing questionable features: tortured writing style, citation of non-existent literature, and unacknowledged image reuse. Surprisingly, some websites offer to rewrite texts for free, generating gobbledegook full of tortured phrases. We believe some authors used rewritten texts to pad their manuscripts. We wish to raise the awareness on publications containing such questionable AI-generated or rewritten texts that passed (poor) peer review. Deception with synthetic texts threatens the integrity of the scientific literature.

• The paper identifies AI-generated 'tortured phrases' as a signature issue undermining clarity in scientific writing.
• It highlights case studies from journals like 'Microprocessors and Microsystems', where fast review cycles hint at compromised editorial standards.
• The study employs GPT detectors and examines article spinning tools to expose trends that erode academic publishing integrity.

Analysis of "Tortured phrases: A dubious writing style emerging in science"

The paper by Cabanac et al., entitled "Tortured phrases: A dubious writing style emerging in science," addresses the critical issue of nonsensical and AI-generated scientific texts infiltrating reputable academic journals. The authors introduce the term "tortured phrases" to describe unexpected word substitutions that replace standard scientific terminology, such as "counterfeit consciousness" instead of "artificial intelligence."

Summary of Findings

The paper documents several concerning trends in scientific literature:

1. Prevalence of Tortured Phrases: Through the analysis of various publications, the authors identify numerous instances of tortured phrases, hypothesizing that these result from advanced LLMs or translation tools that misinterpret standard phrases.
2. Case Study - 'Microprocessors and Microsystems' Journal: A focused examination of the journal 'Microprocessors and Microsystems' reveals a high concentration of articles flagged with synthetic abstracts. The study notes irregularities such as abrupt changes in editorial timelines and short durations between submission and acceptance, often less than 30 days, which imply deficient editorial standards.
3. Artificial Text Generation: They speculate that advanced NLP models, possibly including translation and summarization tools, are being improperly utilized, leading to the dissemination of nonsensical texts. The authors employed a GPT detector to assess the probability of texts being AI-generated, revealing a substantial presence of high-probability synthetic texts in recent publications.
4. Potential Sources and Techniques: The discussion introduces article spinning tools like Spinbot, which replace words with synonyms, as possible contributors to the propagation of tortured phrases. These tools highlight the ease with which authors can produce syntactically correct but semantically flawed scientific papers.

Implications and Future Directions

The implications of this research are profound, raising questions about the integrity of scientific publishing and the potential erosion of trust in academic literature. The paper suggests additional monitoring and investigation into editorial processes and emphasizes the need for developing robust methods to detect and prevent the infiltration of AI-generated or otherwise manipulated texts in scientific publications.

Future research could explore:

• Developing automated tools for detecting tortured phrases across broader scientific disciplines.
• Investigating potential connections between detected patterns and specific article-producing entities (such as paper mills).
• Establishing editorial guidelines for the inclusion of AI-generated content in scientific literature.

In conclusion, the paper serves as a call to action for the scientific community to safeguard the integrity of the academic publishing process by increasing awareness and collaborating on multi-faceted solutions to identify and address the presence of AI-generated nonsensical texts in scholarly articles.