A recent analysis of academic manuscripts on arXiv has revealed the use of "hidden prompts" within papers submitted for AI-assisted peer review. These prompts, concealed through methods like white text or microscopic fonts, instruct AI reviewers to provide positive evaluations or follow specific frameworks. The practice is considered a questionable research practice, with authors' motivations ranging from naive copying to deliberate manipulation. Publishers like Elsevier and Springer Nature have inconsistent policies on AI use in peer review, highlighting vulnerabilities in academic evaluation processes, including plagiarism detection and citation indexing. AI
IMPACT Exposes vulnerabilities in AI-assisted academic review processes, necessitating new technical screening and policy harmonization.
RANK_REASON The cluster is based on an academic paper detailing a novel method of exploiting AI systems within a specific domain. [lever_c_demoted from research: ic=1 ai=1.0]
- alphaXiv
- arXiv
- CatalyzeX
- DagsHub
- Elsevier
- Gotit.pub
- Hugging Face
- ScienceCast
- Springer Nature
- Zhicheng Lin
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
