Automation of Systematic Reviews with AI: A Case Study

• In Event: Contemporary criminological challleges & qualitative methodology

Abstract

Systematic reviews are a fundamental tool in criminological research, as they allow for the rigorous collection and synthesis of available scientific evidence on a given phenomenon. However, this process can be highly demanding in terms of time and resources. This study explores the potential of Artificial Intelligence (AI) to optimize and automate key stages of a systematic review, using the analysis of polarization on social media in empirical research as a case study.

To conduct this study, we followed the PRISMA protocol, combining traditional review methods with AI tools at various stages of the process. AI was used to automatically identify responses to research questions within each article. Additionally, an automated tool for assessing methodological quality (MMAT) was applied to evaluate the rigor of the studies included in the sample. The results show that automation significantly accelerated the identification of relevant information, as well as allowing the classification of articles based on their methodological quality. However, each result was manually reviewed to ensure the reliability of the analysis.

This study highlights the potential of AI to optimize methodological protocols within criminology, acting as a complementary resource that enhances efficiency and accuracy in the synthesis of scientific knowledge, without replacing human supervision.

Author

• Sara Sampayo Sande, Crímina Center - Miguel Hernández University of Elche