AI in Literature Reviews: a survey of current and emerging methods

Abstract

AI-assisted literature review tools have significantly enhanced researchers' productivity by streamlining the review process and reducing time consumption. Although traditional tools remain widely used, a new generation of tools leveraging state-of-the-art methods, including large language models (LLMs), is gaining popularity. However, LLMs face challenges such as hallucinations, which affect their reliability and accuracy. To address this, solutions such as knowledge augmentation are being explored. Additionally, combining knowledge-augmented LLMs with agentic frameworks has shown promise in improving their performance, making them more reliable and effective for literature review tasks.