Testing AI-generated judicial personas

This project explores the growing use of artificial intelligence in the legal profession to create “personas” of judges. These systems are trained on past decisions, legal reasoning, and perceived judicial attitudes to simulate how a judge might respond to a case. Lawyers can then use these simulated responses to test arguments and refine litigation strategy.

The project examines how these tools work in practice, what assumptions they rely on, and how accurate or useful they truly are. It also considers broader concerns, such as whether these systems oversimplify judicial reasoning, reinforce bias, or create false confidence in predicting legal outcomes. By working directly with legal professionals, this research will test the real-world application of these technologies rather than treating them as purely theoretical tools.

For this project you will work with the supervisor on both conceptual research and practical experimentation. Your work will include:

• Conducting desk-based research on how AI is currently used in legal practice, with a focus on predictive tools and judicial modelling.
• Reviewing academic literature, industry reports, and examples of tools that attempt to simulate judicial reasoning or outcomes.
• Assisting in the design of an academic experiment to test how AI-generated judicial personas can be created and used.
• Collaborating with a Brisbane-based law firm and a magistrate to help trial the creation of a judicial persona based on real materials (e.g. case law, decisions, and reasoning patterns).
• Observing and documenting how lawyers interact with these tools, including how they interpret and rely on the outputs.
• Identifying limitations, risks, and complications—such as inaccuracies, over-reliance, or unintended consequences in legal strategy.
• Assisting in organising findings and supporting analysis of how these tools affect legal thinking and practice.

This role combines traditional legal research with exposure to emerging technologies and direct engagement with practitioners.

The project will result in an academic journal article analysing the use of AI-generated judicial personas in legal practice.