@ the IEEE International Conference on Responsible Artificial Intelligence, Melbourne, Australia, 3-September

Abstract:
Technology-facilitated abuse is a recognised form of family violence, yet prevention programs have consistently failed to reach refugee and migrant communities. This is not because the need is absent, but because programs are designed and delivered outside community structures. This paper proposes a community co-design model in which AI literacy is embedded within practical digital literacy training as a vehicle for family violence prevention in Afghan communities in Melbourne. Rather than naming family violence directly — which deters participation — the model uses everyday digital safety concerns as the entry point. Trust and relational belonging are the conditions that make digital safety possible — and that make vulnerability visible before harm occurs. The model operates at two levels: basic digital safety for community members, and AI literacy upskilling for community leaders. Drawing on a Participatory Action Research framework with Bakhtar Community Organisation and the Multicultural Women’s Alliance Against Family Violence (MWAAFV), we argue this model addresses three failures: the detection gap in AI systems that fail to identify family violence in Culturally and Linguistically Diverse (CALD) communities — with fatal consequences; the reach failure of existing prevention programs; and the scalability problem of externally designed interventions. In doing so, it argues that responsible AI requires not only better systems but responsible knowledge production — communities most at risk must be present in generating the knowledge that shapes those systems.

@ Maturitas (in press)

Pisana, B., Stapleton, Deane, F.P., and Beekhuyzen, J.

Objectives: Quality of Life research in Primary Ovarian Insufficiency is limited by the under-representation of women aged ≤25 and the lack of attention to how age at diagnosis shapes outcomes. This study aimed to address these gaps by employing recruitment strategies to increase participation of younger women, while exploring age at diagnosis as a key influence on their experiences.

Study Design: A qualitative design was employed using an online open-ended questionnaire distributed through social media support groups. Data was analysed using thematic analysis.

Results: Participants were 191 women, aged 20-56 years from three English speaking countries. Women described debilitating symptoms affecting sleep, cognition, mood, sexual functioning and relationships, alongside treatment related burdens that disrupted education and employment. Difficulties accessing timely diagnosis and care were reported, citing provider knowledge gaps and limited adolescent-focused diagnostic expertise. A number of women disclosed a history of suicidal ideation, with some attributing it to untreated hormone related symptoms and others to existential loss related to infertility and disrupted life trajectories. Age at onset shaped identity disruption; adolescence-onset was associated with developmental stalling. Women approaching midlife described a ‘care cliff’ where they felt abandoned by services. Across groups, interactions with health care providers were central to adjustment and long-term health.  

Conclusions: Reports of suicidal ideation highlighted the need for both optimised hormonal treatment and psychological support.  Participants reported that many clinicians were inadequately prepared to manage the condition. Provider education in Premature Ovarian Insufficiency is essential, particularly to improve timely adolescent diagnosis and prevent irreversible permanent developmental impacts.

Beekhuyzen, Nielsen & von Hellens (2010)

This discusses strategies for coding empirical data based on a doctoral study.

Click here to read.