Evidence behind the AI assessmentWhat EAIPAM Evaluates
EAIPAM treats potential AI-mediated drafting as a signal for closer scholarly review, not as a standalone conclusion. It then evaluates Reference Integrity, Citation-to-Claim Alignment, Scientific Accuracy and Evidence Depth to examine whether the scholarship behind the text is traceable, supported and coherent.
AI-generated scholarly text can appear fluent and authoritative while presenting fabricated or inaccurate references, attaching genuine sources to unsupported claims, combining incompatible scientific concepts or producing broad discussions without adequate quantitative or comparative evidence. Language-pattern detection alone cannot establish whether these substantive scholarly problems are present.
Reference Integrity
Reference Integrity verifies whether cited publications are authentic and whether bibliographic details such as the DOI, title, authors, journal, year, volume and pagination are accurate and complete.
Importance in scholarly literature
Accurate references allow readers, reviewers and editors to trace the evidence supporting a manuscript. Fabricated, incomplete, duplicated or incorrectly attributed references weaken scholarly credibility and prevent independent verification.
AI-generated content frequently invents publications, combines bibliographic details from different sources, assigns incorrect DOIs, misattributes authors and produces references that do not correspond to identifiable scholarly records.