Applications of Artificial Intelligence in Intimate Partner Violence Research: A Bibliometric Review and Systematic Analysis

Background:
Artificial intelligence (AI) has increasingly been applied across public health domains, including the study of intimate partner violence (IPV). However, the scope, focus, and equity implications of these applications remain underexplored. This study conducted a bibliometric and systematic review to examine how AI technologies have been utilized in IPV-related research over the past decade, with particular attention to gender representation, research populations, and regional paradigms.

Methods:
A comprehensive literature search was conducted across four databases—Web of Science, PubMed, EBSCO, and Cochrane Library—identifying a total of 528 records. After removing 150 duplicates, 378 titles and abstracts were screened. Following exclusion criteria such as non-peer-reviewed content, outdated publication years, and lack of AI relevance, 63 articles were included in the final analysis. The review focused on studies employing AI techniques (e.g., machine learning, deep learning) in the context of IPV, guided by PRISMA protocols. Articles were categorized based on research focus, population (survivors vs. perpetrators), gender representation, and regional origin.

Results:
Gender representation within the included studies showed substantial imbalance: 12 studies explicitly focused on female survivors, while only one study emphasized male experiences, primarily highlighting underreporting among men. Additionally, five studies concentrated on perpetrators of IPV, with minimal attention to gender differentiation, instead linking findings to broader criminological patterns. Regional comparisons revealed distinct research paradigms: studies from Asian contexts (notably China and India) often explored IPV through cultural and relational lenses, utilizing survey-based methodologies, whereas U.S.-based studies more frequently adopted clinical or criminal justice data and emphasized predictive modeling and technical rigor.

Conclusions:
The application of AI in IPV research is growing but remains uneven in terms of gender representation, population focus, and regional epistemologies. Current research predominantly centers on female survivors and Western contexts, potentially limiting the global applicability of findings. There is a critical need to diversify AI-driven IPV research by incorporating culturally sensitive methodologies, expanding representation of male survivors and perpetrators, and promoting equity-oriented frameworks. Addressing these gaps will enhance the relevance and ethical implementation of AI in IPV prevention and intervention efforts.