Chatbot guides women through post-prison challenges

Most women leaving prison face profound disadvantages and rarely have access to the resources needed to settle back into the community. Seemingly simple tasks such as obtaining replacement identification documents or opening a bank account become tangled in complexities.

Now researchers at the University of South Australia are co-designing a chatbot to help formerly incarcerated women re-establish their lives on the outside, and reduce the risk of them returning to prison.

Led by a team of UniSA researchers in collaboration with advocacy group Seeds of Affinity, the tech-based solution aims to help women access trusted and authentic information gathered by other women who have navigated the same complex social services.

In 2023, almost 42,000 people were imprisoned in Australia, 3168 of them female. According to the Australian Bureau of Statistics, the number of female prisoners increased by 6% last year.

UniSA researcher and senior lecturer in social work Dr. Michele Jarldorn says inadequate support for women leaving prison often sets them up to fail, with a high percentage of women back behind bars within two years of their release.

“Women who have been to prison are among the most disadvantaged groups in society. Despite having completed their prison sentence, they’re rarely welcomed back into the community and for many, the only friendships they have are those made in prison,” she says.

“Upon release, women are incredibly vulnerable to homelessness and rather than experiencing release as freedom, it’s a scramble for support, resources and survival. The rates of suicide, poor mental and physical health and drug overdose are much higher than the general population. These women are unlikely to seek help from a mental health provider and if they do, they’re placed on a waiting list as demand for services outweighs capacity in Australia.”

Researchers are developing a prototype chatbot called LindaBot, a piece of software designed to process and resemble human conversation. Chatbots are commonly used on company websites to provide customer services and deal with a high volume of inquiries, at any time of the day or night. The information embedded within this chatbot has been sourced, created and developed by formerly incarcerated women who are working with researchers on the project.

Designed for access on mobile phones, LindaBot will be able to provide information and help with tasks that, on the surface seem simple but, are incredibly complex.

Game designer and UniSA lecturer of games design and digital media Dr. Susannah Emery says the design of the chatbot must ensure the information communicated doesn’t create confusion or frustration in users. Providing emotional support is also a focus.

“When finished, LindaBot will ask the user if they are currently experiencing a crisis or serious mental health concern and if they are, the chatbot provides contact details for 24-hour crisis response services,” she says.

“The user is also asked if they need to speak to a ‘real person’ and if so, they’re invited to speak to someone at Seeds. Women are rarely asked how they are doing or offered emotional support, so with LindaBot, users are asked if they would like to receive daily check-in messages. They can also be sent something that other women from Seeds have found helpful in their journeys post-release, such as videos, sound files or links to music or other resources.”

Researchers are hoping to launch the finished product by early 2025.

Dr. Jarldorn says once the project is complete and LindaBot is formally launched, they will simultaneously begin a campaign to seek donations of second-hand phones that women can use post-release.

Seeds of Affinity recently recruited people into paid positions to establish a LindaBot advisory group, assisting in workshop facilitation, data analysis and working alongside the tech development team.

“It costs more than $115,000 a year to imprison one woman, so ongoing funding and support to maintain LindaBot is a small investment with a significant return,” Dr. Jarldorn says.

“This project seeks to involve and empower a community of women with very little social power but so much knowledge and lived experience. LindaBot has the potential to be incredibly meaningful for a cohort that tends to be marginalized and socially excluded, and usually told what they will get rather than asked that they want or need.”

Their current research is published in the International Journal of Communication.