Title: System for systematic literature review using multiple AI agents: Concept and an empirical evaluation

Addressing the labor-intensive nature of Systematic Literature Reviews (SLRs), this paper introduces a multi-AI agent model to automate the process. By utilizing Large Language Models (LLMs), the model performs a range of functions from generating precise search strings to summarizing abstracts and conducting in-depth analysis on selected papers, all tailored to predefined research questions. With empirical evaluation conducted by software engineering researchers, the proposed model shows great promise in transforming traditional literature review methods.

• SLRs are foundational but arduous, often requiring manual labor.
• The multi-AI agent model automates the process, using LLMs to increase the efficiency and accuracy of SLRs.
• The agents handle every phase of the SLR: search string generation, inclusive and exclusive filtering, abstract summarization, and analysis.

The implications of this research are profound for the academic community, as it can significantly streamline the SLR process, allowing for more robust and quicker synthesis of existing studies. This AI-based approach represents a major leap forward in the methodology of evidence-based research, where the precision of AI complements the expertise of human researchers.