Bridging the Gap: Leveraging AI to Prepare Social Work Students for Practice with IPV Survivors

Background and Purpose: Intimate partner violence (IPV) is highly prevalent, impacting an estimated 41% of women and 26% of men in the United States over their lifetime, and is associated with significant physical, emotional, and social consequences. Social workers come into contact with IPV survivors across multiple settings, yet preparing social work students to safely engage with survivors in ways that are trauma-informed and survivor-centered remains a persistent challenge in social work education. Simulation-based techniques provide structured, interactive opportunities that mirror real-world situations in a low-risk environment for practicing IPV skills without risking harm or re-traumatization. The purpose of this mixed-methods study is to identify high-priority IPV learning competencies and develop an empirically-derived Artificial Intelligence powered Survivor Simulation Bot (AI-SSB) for use across a variety of undergraduate and graduate courses.

Methods: Quantitative pre-test/post-test survey data were collected from students before and after enrollment in three undergraduate and graduate IPV courses. Paired-samples t-tests were used to examine how the courses impacted IPV preparedness, myth acceptance, and relational empathy, providing insight into student readiness and gaps that could be addressed through the AI-SSB. Qualitative data were gathered from students, faculty, and advocates who participated in three semi-structured meetings to understand the specific competencies and skills that the AI-SSB could target in training future social workers and domestic violence advocates.

Results: Survey participants (N = 22) were female (100%) with a mean age of 32 (SD = 11.5). Results indicated that students' preparedness to engage in IPV practice (t (21) = -5.68, p < .001) and relational empathy (t (21) = -4.96, p < .001) increased from pre- to post-test. However, there was no significant change in IPV myth acceptance from pre- to post-test. Qualitative data gathered from advocates (n = 6), social work faculty (n = 4), and students (n = 3) revealed that the AI-SSB could strengthen preparedness through repeated, safe practice opportunities that reflect the varied needs and priorities of survivors. Students further indicated a strong desire to use AI-powered tools to advance learning, noting, “The bot doesn’t seem like extra work...and like something I’d be excited to use, supporting skill development over time.”

Conclusions and Implications: Preliminary findings suggest that an AI-SSB could improve students' preparedness, skills, and sensitivity. The AI-SSB may also provide opportunities to address IPV myth acceptance by increasing student exposure to realistic and dynamic survivor scenarios. Building on these findings, an AI-SSB has been developed for use in social work IPV courses. It is expected that this will provide an interactive, low-risk environment that allows students to practice survivor-centered, trauma-informed, and empowerment based approaches prior to engaging with IPV survivors in real world settings. Incorporating technology-based tools and applied learning approaches in social work education has the potential to enhance student confidence, preparedness, and applied skills while reducing the potential for harm inherent in practicing social work skills with IPV survivors.