Between Similarity and Synthetic Authorship: Reconfigurations of Academic Integrity in the Era of Computational Cognitive Systems

Abstract

Background:
The rapid expansion of Artificial Intelligence (AI) in higher education has reshaped teaching, assessment, and academic writing practices, exposing the limitations of traditional plagiarism detection models. At the at the same time, the emergence of advanced algorithmic systems and generative AI has intensified ethical, pedagogical, and institutional debates concerning authorship, academic integrity, and fair assessment.
Objective:
To critically analyze the impact of Artificial Intelligence on plagiarism detection in higher education, considering its technical, ethical, and pedagogical implications.
Method:
A systematic literature review was conducted following the PRISMA protocol, using international databases including Scopus, Web of Science, ERIC, IEEE Xplore, and Google Scholar. A total of 963 records were identified, of which 18 studies met the eligibility criteria and were included in the final analysis.
Results:
The findings reveal growing reliance on AI- based plagiarism detection systems, alongside persistent technical limitations such as overreliance on similarity scores, algorithmic bias, false positives, and reduced effectiveness in identifying AI- generated texts. The results also highlight significant effects on teaching and assessment practices, particularly when automated outputs are applied without adequate pedagogical mediation.
Conclusions:
Artificial Intelligence reshapes plagiarism detection practices but does not replace contextualized human judgment. Its responsible use requires clear institutional policies, strengthened academic and ethical literacy, and pedagogical approaches that prioritize formative processes over punitive measures, particularly within Global South contexts.