Phantom and Truth: Identification and Avoidance Strategies for Misinformation and Hallucinated Citations in AI-Generated Academic Content

Abstract

With the widespread application of generative artificial intelligence in academic writing, misinformation and hallucinated citations have frequently appeared in AI-generated papers, books, and articles, posing a severe challenge to academic integrity. This paper systematically categorizes the typical characteristics of misinformation and the manifestations of hallucinated citations in AI-generated content. It proposes identification methods from three dimensions: source verification, logical testing, and technical tool assistance, and constructs avoidance strategies across four levels: author self-discipline, academic norms, technical prevention, and institutional guarantees. The study concludes that addressing the academic integrity crisis triggered by AI requires multi-party collaboration to find a balance between technical empowerment and academic bottom lines, ultimately ensuring the reliability of academic research.