- Browse/Search the Program
Search
Program Calendar
Browse By Day
Browse By Room
- Information Menu
Search Tips
Virtual Exhibit Hall
- Navigation and Settings Menu
Personal Schedule
Sign In
Applying GPT Models to Extract Criminal Data from Intimate Partner Femicide Court Rulings: Preliminary Results
Fri, September 5, 8:00 to 9:15am, Deree | JSB Library, Floor: Main level, JSB Library Conference Room [LCR]Abstract
Criminal court rulings contain critical information about crimes, offenders and victims, making them an essential resource for understanding severe forms of violence. Curren literature has evidenced that taking this information into consideration contributes to the prediction and prevention of intimate partner femicide (IPF), the most frequent form of intimate partner homicide. Recent advancements in artificial intelligence (AI), particularly in foundational transformer-based models, have significantly enhanced what is denominated natural language processing (NLP). One the most renowned model worldwide is GPT (Generative Pre-trained Transformer) made by OpenAI. Many studies have explored AI and NLP applications for summarizing and extracting data from legal texts. However, there is still a gap in their application to criminal court rulings and especially to cases of severe violence such as IPF. The current study aims to explore how these new models contribute to a deeper and more comprehensive understanding of IPF cases. We used a GPT model to summarize 458 Spanish criminal court rulings on IPF. Using manually codified variables, we evaluated the performance of GPT models in extracting key criminological data, such as criminal history, sanctions, situations, and characteristics of partner violence. These results demonstrate that several variables can be extracted with high performance and efficiency, highlighting the potential of these technologies for large-scale legal text analysis. This study contributes to the development of automated, efficient methods for analyzing legal texts, facilitating the work of researchers and professionals in the justice system and contributing to their final goal of saving lives. By providing structured relevant information from court rulings, this approach can support crime prevention strategies and inform public policies aimed at addressing one of the most serious manifestations of human violence.