Plagiarism, and Hallucinated Citations [Han Seung-hoon Column]

The process of peer review for academic journals, a routine task for researchers, involves rigorously assessing manuscripts for adherence to format, the integrity of data, and relevance to the journal's scope. This evaluation is typically conducted by two to three anonymous experts who collaborate with authors to enhance scholarly work.

“

The term 'plagiarism' is used too lightly in South Korean society. This word is often misused as a tool to attack public officials or celebrities who obtained degrees easily for career management without professional academic training.

While the author generally advocates for a lenient approach to publication, acknowledging that not all papers are groundbreaking, certain submissions must be rejected. The most egregious example is plagiarism, a term the author feels is misused in South Korea. Unlike minor citation errors, plagiarism is defined as intentionally stealing intellectual property, and its victims and deliberate nature necessitate severe consequences.

Recent years have seen the emergence of a new category of problematic papers: those containing "hallucinated citations." This phenomenon, where fabricated references are cited, is alarmingly widespread. Research indicates that in 2025 alone, nearly 150,000 false citations were identified, with a significant surge occurring after the widespread adoption of generative AI in late 2024. While AI tools offer efficiency in literature review, their "hallucination" tendency, if unchecked by thorough author review, leads to the proliferation of fake references.

“

Therefore, the term plagiarism should only be used when there is a victim and the perpetrator intentionally presents part or all of the achievement as their own.

The victims of plagiarism have their work stolen and attributed to others. In contrast, victims of hallucinated citations are cited for work they never produced, and in cases where the cited author doesn't exist, the victim is even more ambiguous. While technological advancements are a factor, the issue of citing unread or misrepresented sources has existed historically. However, the ease with which AI can generate fake references poses a severe strain on academic verification systems. The fundamental responsibility lies with researchers to own their work and uphold the traditional function of scholarship in expanding human knowledge, even in the age of AI.

“

The scale of this phenomenon, called hallucinated citation, is shocking.