Slashdot: AI Tool Detects LLM-Generated Text in Research Papers and Peer Reviews

Sep 19, 2025

—

Source URL: https://slashdot.org/story/25/09/19/1750226/ai-tool-detects-llm-generated-text-in-research-papers-and-peer-reviews
Source: Slashdot
Title: AI Tool Detects LLM-Generated Text in Research Papers and Peer Reviews

Feedly Summary:

AI Summary and Description: Yes

Summary: Recent analysis conducted by the American Association for Cancer Research (AACR) reveals a significant increase in AI-generated text within academic submissions, particularly highlighting concerns about disclosure practices among authors. This development raises essential considerations for information security and compliance in academic and research environments.

Detailed Description: The findings from the AACR’s examination of manuscript submissions highlight critical trends and implications for researchers, publishers, and security professionals.

– **Increase in AI-Generated Text**:
– The AACR identified that 23% of abstracts and 5% of peer review reports contained text likely produced by large language models (LLMs).

– **Authors’ Disclosure Practices**:
– Despite a requirement for authors to disclose AI usage in their submissions, fewer than 25% complied with this mandate, raising ethical and transparency issues.

– **Screening Process**:
– The AACR utilized an AI tool from Pangram Labs to analyze a substantial volume of academic submissions, showing a marked increase in flagged AI-generated content from late 2022, coinciding with ChatGPT’s public release.

– **Implications for Academic Integrity and Compliance**:
– The surge in AI-generated text may challenge traditional standards of authorship and originality.
– Publishers and academic institutions may need to implement stricter compliance and governance measures to ensure ethical use of AI tools.

– **Security and Privacy Considerations**:
– The use of AI in preparing scholarly work necessitates careful consideration of how AI tools manage and process sensitive data, particularly in research areas involving personal or confidential information.

In conclusion, this analysis presents significant insights for professionals involved in AI ethics, information security, and academic compliance, urging a reevaluation of policies governing the use of AI in scholarly communications and the implications it carries for governance and authorship standards in research publications.

1 2 3 5 7 a abstract academic institutions academic integrity Act age AI AI Ethics AI tool AI tools AI-generated content All America American analysis and anti Arch art as at ated authors authorship by C care CERN challenge chat ChatGPT CI CIA co communication compliance compliance and governance concerns confidential information content critical D data de development disclosure DoT e end environment environments ethical ethical use ethical use of AI Ethics evaluation for g Gen generated Generated Content Go governance governance measures GPT gs H high Highlight http HTTPS implications in information information security insights institutions integrity io Iron issue ite k l language language model language models large large language model large language models Large Language Models (LLMs) led Li llm llms lm M man measures mission Mode model models N nation o of on ons OPM originality ory oS out over paper Peer Review per policies practices pre privacy privacy considerations pro process professionals ps public publications publishers Q R Raise raising rate RCE re release report reports research research environments research papers researchers review reviews Ro s scholarly communication Screening Process search sec security security and compliance security professionals sensitive data side Sig SoC source SSE SSO standards T ted text the to tool tools Tor TP transparency transparency issues trends UI US usage use uth V val Valuation Wi x z