91ɫ��Ƭ

As AI-enabled abuse grows, examining perpetrators tactics and developing prevention strategies is critical.

A 91ɫ��Ƭ Sydney-led research project will investigate how generative artificial intelligence is being used to facilitate coercive control and gender-based violence against women and children.

The project, led by Dr Delanie Woodlock from 91ɫ��Ƭ’s Gendered Violence Research Network (GVRN), is one of 12 to receive funding through the final round of the .

It is being delivered in partnership with , Australia’s peak body for women’s domestic and family violence services, and Professor Donna Chung from Curtin University, with research support from 91ɫ��Ƭ Sydney’s Social Policy Research Centre.

“Domestic and family violence services warned two decades ago that technology would be used to facilitate coercive control. They are now warning of a new wave of harm enabled by generative AI,” said Dr Woodlock.

Public and academic concern has largely focused on deepfakes. However, AI can also be used to impersonate people and fabricate ‘evidence’ designed to discredit women, manipulate their relationships and undermine their credibility.

of domestic and family violence practitioners conducted by the research team found that 74% had seen perpetrators use AI-generated threats or impersonation tools to abuse women.

Dr Woodlock said the project would examine how these tactics were being used and identify ways to prevent technology from becoming a tool of abuse.

“Generative AI is making it easier for perpetrators to control, impersonate and intimidate women. We have a narrow window to act before these tactics become normalised,” Dr Woodlock said.

“This project will show us what drives AI-facilitated abuse and what works to stop it. The goal is to build safety into these tools from the start, so the burden of staying safe does not fall on survivors.”

The project will:

• interview victim-survivors and conduct focus groups with frontline practitioners to document the tactics being used and their effects on women and children
• conduct a nationally representative survey of Australian men to identify the attitudes that encourage or deter men from using AI to intimidate, threaten and harass women and children
• develop Safety by Design guidelines for technology companies and government, co-designed with people with lived experience and grounded in the eSafety Commissioner’s Safety by Design principles, and
• bring together technology companies, victim-survivors and the domestic and family violence sector for an industry roundtable.

“The project does not focus on changing the behaviour of victim-survivors. Instead, it will examine the systems perpetrators exploit and the social attitudes that can normalise coercive control,” explains Dr Woodlock.

“Its goal is to shift the burden of safety from survivors to the technologies being developed and the systems that govern their use.”

Professor Jan Breckenridge, co-convener of the Gendered Violence Research Network and Head of the School of Social Sciences at 91ɫ��Ƭ Arts, Design & Architecture, said the project addressed an urgent threat.

“This project addresses a critical and rapidly emerging threat, examining how generative AI can be misused to facilitate coercive control and gender-based violence,” Professor Breckenridge said.

“By developing evidence-based recommendations for industry, government and the technology sector, Dr Woodlock’s project will help to shape safer technologies and better outcomes for women, children and communities.”