From Data to Action: How Artificial Intelligence (AI) Can Help Combat Gender-Based Violence

Author:

Elisabetta Aldrovandi - Attorney & External Consultant to the Femicide Commission of the Chamber of Deputies & Professor of Criminology and Victimology at LIMEC-SSML Milan

Abstract

Artificial Intelligence (AI) has been an integral part of our daily lives for some time now, but it is only in recent years that we have started to extensively discuss and explore its characteristics and potential. 

This exploration is relevant both in the context of combating gender-based violence—expression that is comprehensive of all forms of violence, whether psychological, physical, or even sexual, manifesting through legally significant acts such as stalking, sexual assault, and the worst act of all, murder—and in addressing domestic violence, which encompasses these types of violence within the family or an intimate relationship.

AI-Powered Solutions for a Safer World: Addressing Gender Violence

Special attention has to be given to the systems of sentiment analysis, a term referring to the process of analyzing digital text to identify the emotional tone of the message as positive, negative, or neutral. This analysis employs methodologies based on Deep Learning architectures (a subset of Artificial Intelligence). 

The accuracy of sentiment and emotion analysis results is ensured not only by the intrinsic quality of the methodology but also by the training phase activities, employing Active Learning techniques.

Starting from a reference topic, it is possible to gather a series of sentiment responses (e.g. positive, negative or neutral evaluations) and emotions (e.g. joy, surprise, love, sadness, anger or neutral) and, depending on the reference topic selected, the accuracy of those classifications is ensured by the employment of appropriate reference elements that are tailored to the specific needs of the context. In our case, gender-based and domestic violence.

By leveraging AI and its capabilities, we aim to develop experimental statistics to provide new information and fill a significant informational gap, by offering valuable unprecedented insights into the public sentiment (mood) surrounding key events related to gender-based and domestic violence. 

The goal is to understand how social media contents dealing with these issues are conveyed through positivity or negativity indexes and to observe how the phenomena themselves have evolved, incorporating new digital expressions that often benefit from (not only physical) shielding, making it more challenging to identify the perpetrators.

A concrete example

A concrete example of what has just been stated above, is represented by the algorithms developed in partnership with Binoocle, a start-up from Florence, that analyzed a real case: I am talking about the reactions to an interview with Alessandra Pauncz, President of the National Network of Centers for Perpetrators of Violence, conducted for the website Storie di Firenze, with the support of Cesvot.

The interview, posted on social media, gained 347 comments, 1.351 interactions and 596 clicks.

The AI examined user demographics, their views on the topic, their expressions and the responses were diverse: to fight gender-based violence, 14,9% of commenters called for social change, especially for significant changes in societal, cultural, and value paradigms; another 14,9% would like a legislative and punitive reform that would aggravate the sentences; 11,9% of the users highlighted the direct responsibility of men, emphasizing the need for changes in attitudes and behaviours; other users (10.5%) acknowledged the role of media and politics in influencing perceptions and behaviours related to gender violence.

Conclusion

Positive use of AI, beneficial to the community as a whole, is represented by the careful attention on violent phenomena and the monitoring of the sentiment related to topical issues, specific events and legislation initiatives aimed at preventing and fighting gender- based and domestic violence, to foster respect and empathy towards one another.

New manifestations of violence require developing new analytical tools capable of detecting, preventing, and fighting it. Paradoxically, it is through the integration of AI that society gains a higher potential to cultivate sensitivity and a more human-centric approach.

Biography of the Guest Expert

Ms. Elisabetta Aldrovandi, an Attorney with a long history of commitment to the protection of crime victims, has been practicing Law since 2000 and she is renowned for her dedication to advancing tangible social justice.

Since 2017, as the Co-founder and President of the National Support Victims Observatory, Attorney Aldrovandi has focused on formulating legislative proposals to effectively safeguard crime victims through prevention and punitive as well as rehabilitative measures. 

Her participation in hearings before the Justice Commissions of the Chamber of Deputies and the Senate, and her role as an External Consultant to the bicameral commission of inquiry on Femicide at the Chamber of Deputies, are particularly noteworthy in this regard.

Furthermore, since 2019, Attorney Aldrovandi has been a Professor of Criminology and Victimology at LIMEC-SSML, the Higher School for Language Mediators in Milan. 

From 2019 to 2024, she served as the Guarantor for the protection of crime victims for the Lombardy Region. 

She has been included among the 100 most influential women of 2019 by the weekly magazine F. and She is a regular guest on national television networks as an expert on crime cases.