From Verification Burden to Trusted Collaboration: Design Goals for LLM-Assisted Literature Reviews

Although numerous studies have examined how researchers use Large Language Models (LLMs) in academic writing, their concrete implementation, limitations, and design challenges in the literature review process remain underexplored. In this paper, we present findings from a user study with researchers across multiple disciplines that characterize current practices, benefits, and pain points when using LLMs to investigate related work. We identified three recurring gaps: (i) lack of trust in outputs, (ii) a persistent verification burden, and (iii) fragmentation across tools. To address these gaps, we propose six design goals and a high-level framework that operationalizes them through improved visualization of related papers, verification at every step, and human-feedback alignment with generation-guided explanations