Researchers receive award for AI tool making community legal advice more accessible

L-R: Dr Liming Zhu (AFR AI Awards Judge), Associate Professor Jey Han Lau, Dr Mel Mistica, Dr Kemal Kurniawan, Rose Hyland (Senior Project Lead, Justice Connect) and Tom O'Doherty (Head of Innovation and Connections, Justice Connect).
L-R: Dr Liming Zhu (AFR AI Awards Judge), Associate Professor Jey Han Lau, Dr Mel Mistica, Dr Kemal Kurniawan, Rose Hyland (Senior Project Lead, Justice Connect) and Tom O'Doherty (Head of Innovation and Connections, Justice Connect).

Researchers from the Faculty of Engineering and Information Technology have won the community engagement award at the 2026 Financial Review AI Awards for an AI tool making community legal advice more accessible.

The tool was developed in partnership with community legal centre Justice Connect (JC). Rather than dispensing advice to people needing legal assistance, it helps the centre quickly understand and triage the legal questions the client needs answered.

JC provides free legal advice, often to vulnerable communities. Prior to the AI tool, determining what type of legal advice people need, and triaging them accordingly, was resource intensive for the pro-bono service, taking up valuable time lawyers could have spent providing legal advice instead.

Dr Kemal Kurniawan, who led the team of researchers, said they were incredibly proud of their work to support JC.

“The tool we created is already having a major impact on the community and means JC’s lawyers can help more people,” Dr Kurniawan said.

JC initially launched an online questionnaire in 2017 that asked technical legal questions, but people often abandoned it before completion.

By 2020, the partnership between JC and CIS researchers began to address this.

Dr Kurniawan, Dr Mel Mistica, Professor Tim Baldwin and Associate Professor Jey Han Lau developed an AI tool which has users explain legal problems in their own words. The system then sorts their issues into 32 legal categories and triages them with notes, ready for JC lawyers.

The AI was trained by JC lawyers, who were assigned sample texts from diverse groups of people to sort into the categories.

It now accurately classifies the legal area of a user’s problem around 75 per cent of the time, compared with roughly 35 per cent accuracy from the everyday person, reducing abandoned applications by 48 per cent and increasing legal service delivery by 10 per cent.

See the full list of 2026 Financial Review AI Awards winners.