Periodic Reporting for period 1 - IMPROVE (Improving Access to Services for Victims of Domestic Violence by Accelerating Change in Frontline Responder Organisations)
Reporting period: 2022-10-01 to 2024-03-31
Domestic violence (DV) and violence in close relationships still persist as serious forms of violence and crime in Europe. In the field of prevention and fight against DV and gender-based violence against women, the EU regards the Istanbul Convention as the benchmark for international standards. Victims of DV should have an easy and secure access to justice and necessary services, assistance and support. The real success (or failure) of the policies, action plans and legislation are tested in their implementation and enforcement.
IMPROVE aims to increase DV reporting and improve access to DV services for under-served, marginalised victims, by providing a novel entry point to existing frontline responder networks by means of an effective conversational AI Chatbot.
IMPROVE aims to achieve this objective, first, by identifying the individual and structural barriers of victim-survivors to the existing service network through desk research and empirical field research (WP1); second, by the technical development of the conversational AI Chatbot integrating the identified needs (WP2); and third, by piloting and validation of the conversational AI Chatbot (WP2).
Furthermore, IMPROVE aims to accelerate DV policy implementation using a validated common approach for frontline responder organisations and policy makers to reinforce and exploit existing solutions and enablers, while removing barriers to change and overcoming organisational inertia.
IMPROVE has planned to achieve this objective, through problem analysis and solution development: The specific work steps comprises of empirical qualitative research identifying existing challenges as well as positive change factors within DV public policy and FLP organisations and the development of organisational decision support tools spanning across two Work Packages (WP1, WP4).
To complement the service provision for under-served, marginalized victims, IMPROVE not only has devoted effort to widen the access points of victims, but to improve the work performed by the practitioners alike. IMPROVE wants to improve multi-agency DV responses for underserved populations by fostering capacity building of frontline practitioners through scalable, low-threshold, targeted training, raising mutual understanding and bridging sectoral divides.
IMPROVE is to widen the training modules with respect to professions addressed, languages available, and content included: IMPROVE’s multi-agency approach encompasses key sectors crucial for effective DV management.
Complementing all efforts and goals of IMPROVE, it is the overarching and integrating objective to raise awareness among, and empower, underserved European populations to access existing DV services, while increasing FLR organisations’ and policy makers’ knowledge and awareness of underserved populations’ needs.
WP5 draws from the research findings, as well as the development of technical and organisational solutions as well as capacity building materials to inform its strategy and messaging. In that, its achievement is linked to the progress of the work done in all WPs.
IMPROVE aims to increase DV reporting and improve access to DV services for under-served, marginalised victims, by providing a novel entry point to existing frontline responder networks by means of an effective conversational AI Chatbot.
IMPROVE aims to achieve this objective, first, by identifying the individual and structural barriers of victim-survivors to the existing service network through desk research and empirical field research (WP1); second, by the technical development of the conversational AI Chatbot integrating the identified needs (WP2); and third, by piloting and validation of the conversational AI Chatbot (WP2).
Furthermore, IMPROVE aims to accelerate DV policy implementation using a validated common approach for frontline responder organisations and policy makers to reinforce and exploit existing solutions and enablers, while removing barriers to change and overcoming organisational inertia.
IMPROVE has planned to achieve this objective, through problem analysis and solution development: The specific work steps comprises of empirical qualitative research identifying existing challenges as well as positive change factors within DV public policy and FLP organisations and the development of organisational decision support tools spanning across two Work Packages (WP1, WP4).
To complement the service provision for under-served, marginalized victims, IMPROVE not only has devoted effort to widen the access points of victims, but to improve the work performed by the practitioners alike. IMPROVE wants to improve multi-agency DV responses for underserved populations by fostering capacity building of frontline practitioners through scalable, low-threshold, targeted training, raising mutual understanding and bridging sectoral divides.
IMPROVE is to widen the training modules with respect to professions addressed, languages available, and content included: IMPROVE’s multi-agency approach encompasses key sectors crucial for effective DV management.
Complementing all efforts and goals of IMPROVE, it is the overarching and integrating objective to raise awareness among, and empower, underserved European populations to access existing DV services, while increasing FLR organisations’ and policy makers’ knowledge and awareness of underserved populations’ needs.
WP5 draws from the research findings, as well as the development of technical and organisational solutions as well as capacity building materials to inform its strategy and messaging. In that, its achievement is linked to the progress of the work done in all WPs.
With respect to victims’ needs, the problem identification commenced with the analysis of the national level scientific literature and data sources to identify internal and external impediments to seek help experienced by victim-survivors, as well as existing gaps in information and knowledge, which was complemented by extensive empirical field research: First, we conducted 132 interviews with victim-survivors who already found their way into the help system learning from their experiences about challenges and drivers of successfully receiving help.
Building on this knowledge, WE’s therapists developed a Requirement Specification Report for the AI tool. This report outlines the tool's overall functionality, priority categories, roles of the content management system (CMS) and Chatbot, and the high-level architecture for data collection, sharing, and storage – resulting in a clickable mock-up of the AI tool.
Against the background of the systemic and structural barriers experienced by victim-survivors (T1.1) T1.3 analysed existing challenges in service provision (accessibility, quality of service, operational and procedural gaps) by conducting expert interviews with frontline practitioners of different sectors. Going beyond problem definition, a typology of improvement paths for FLP organisations with respect to implementing international policies and standards identifying positive change factors as well as specific good practices in victim support systems identified in the partner countries was developed.
In the first half of the project, English and German content was published at the IMPROVE training platform. By building on the IMPRODOVA training platform as a foundation and identifying findings from the initial desk and field research (T3.2) its scope expanded to incorporate victim-survivors’ perspectives and needs. Moreover, the incorporation of insights into the challenges faced by frontline responders during crises such as pandemics (T1.4) enriches the training content, enabling professionals to adapt and respond effectively in both normal and exceptional conditions. Through collaborative efforts with stakeholders from partner countries, WWU and partners are currently (T3.3) designing innovative training formats and optimising existing tools to empower frontline responders in combating DV and facilitating victim-survivors’ access to essential services and justice.
Building on this knowledge, WE’s therapists developed a Requirement Specification Report for the AI tool. This report outlines the tool's overall functionality, priority categories, roles of the content management system (CMS) and Chatbot, and the high-level architecture for data collection, sharing, and storage – resulting in a clickable mock-up of the AI tool.
Against the background of the systemic and structural barriers experienced by victim-survivors (T1.1) T1.3 analysed existing challenges in service provision (accessibility, quality of service, operational and procedural gaps) by conducting expert interviews with frontline practitioners of different sectors. Going beyond problem definition, a typology of improvement paths for FLP organisations with respect to implementing international policies and standards identifying positive change factors as well as specific good practices in victim support systems identified in the partner countries was developed.
In the first half of the project, English and German content was published at the IMPROVE training platform. By building on the IMPRODOVA training platform as a foundation and identifying findings from the initial desk and field research (T3.2) its scope expanded to incorporate victim-survivors’ perspectives and needs. Moreover, the incorporation of insights into the challenges faced by frontline responders during crises such as pandemics (T1.4) enriches the training content, enabling professionals to adapt and respond effectively in both normal and exceptional conditions. Through collaborative efforts with stakeholders from partner countries, WWU and partners are currently (T3.3) designing innovative training formats and optimising existing tools to empower frontline responders in combating DV and facilitating victim-survivors’ access to essential services and justice.
1. Wider inclusion of research subjects (victim-survivors, and DV professionals) than envisaged (KPIs exceeded).
2. Additional safety measures with respect to the inclusion of vulnerable groups in research employed (Trauma-sensitive interviewing workshop).
3. After the external user testing of AinoAid™ was conducted, penetration testing (i.e. test whether the chatbot and knowledgebase are easy to hack) by an external cybersecurity company was conducted. The result was Class A, excellent: “The security exceeds Industry Best Practice standards. The overall posture was found to be excellent with only a few low-risk findings identified.”
4. Selection of AinoAid™ by Microsoft Fast Track Service: Microsoft has very strict Responsible AI practices in place. Microsoft also supports WE with its selective Fast Track services. These services include weekly calls during which Microsoft’s own engineers provide consultation and recommendations for privacy, Responsible AI and other relevant matters.
2. Additional safety measures with respect to the inclusion of vulnerable groups in research employed (Trauma-sensitive interviewing workshop).
3. After the external user testing of AinoAid™ was conducted, penetration testing (i.e. test whether the chatbot and knowledgebase are easy to hack) by an external cybersecurity company was conducted. The result was Class A, excellent: “The security exceeds Industry Best Practice standards. The overall posture was found to be excellent with only a few low-risk findings identified.”
4. Selection of AinoAid™ by Microsoft Fast Track Service: Microsoft has very strict Responsible AI practices in place. Microsoft also supports WE with its selective Fast Track services. These services include weekly calls during which Microsoft’s own engineers provide consultation and recommendations for privacy, Responsible AI and other relevant matters.