AI Plagiarism Policy for India in the Age of ChatGPT
DOI:
https://doi.org/10.56042/jipr.v31i6.19742Keywords:
Large language models, ChatGPT, Plagiarism, Turnitin, Plagiarism Software, paraphrasingAbstract
In this age where AI tools such as ChatGPT are used to substitute research and writing, how should plagiarism policies of academic institutions ensure honesty and scientific progress? This paper approaches the question by considering the concept of plagiarism and providing insights into students' progression in adjusting with plagiarism software from the years pre-ChatGPT to years post introduction of ChatGPT. Considering the discovered practicalities, the paper then proposes an AI plagiarism taxonomy. The taxonomy of AI plagiarism classifies the following kinds of plagiarism using AI - verbatim copying plagiarism, uncited paraphrasing plagiarism, AI idea plagiarism, language change plagiarism, misrepresented citation AI plagiarism, and prohibited use of AI plagiarism. The paper then also proposes a two-tiered AI plagiarism policy where the first tier follows the UGC Plagiarism Regulations, 2018. The second tier recognises the need for human oversight and possible software development to flag (1) sentences which are stating a fact or idea that is not common knowledge and where citation is expected (2) sentences which are stating a fact or idea which has been cited; however, the citation is misrepresenting the source and (3) missing acknowledgement of use of AI in idea generation, text generation, paraphrasing and other uses. Therefore, the paper proposes a progressive addition to the traditional understanding of plagiarism as textual similarity to a broader, idea- and citation-based understanding of academic integrity.