Back to Articles
AI and the Courts in 2025

Federal Court of Australia

ENRICHED

Details

Date Published
27 June 2025
Priority Score
4
Australian
Yes
Created
16 June 2026, 04:00 pm

Authors (1)

Description

<a href="https://news.google.com/rss/articles/CBMinwFBVV95cUxPTi0zZGFYaE1JX2xVb25ELTJkbVNhRWhvVE56TjZLcFhtRm1RNFFfbEhTbUxMam4tY096aW01RURXZHpyZDhuYnh1aWFxbkk4Q0pZTWppbmp1UmxFV1JvTWt2Zmtsb1JmTEdUQUQtSnppNUcyc2c0RWRmSl9nOE1rQnVPVVlrbHdjZDEzQU12YlAwV2RqLUtiM09IOUV5aFk?oc=5" target="_blank">AI and the Courts in 2025</a>&nbsp;&nbsp;<font color="#6f6f6f">Federal Court of Australia</font>

Summary

This judicial speech outlines the systemic risks posed by generative AI within the Australian legal system, specifically focusing on the lack of nuanced understanding and reliability in legal reasoning. It highlights how the probabilistic nature of LLMs leads to hallucinations and misinformation, particularly concerning Australian law where training data may be limited. The analysis addresses critical governance issues including data privacy, ethical biases in training sets, and the potential for AI to undermine the integrity of judicial processes. These observations are vital for establishing regulatory safeguards to prevent automated systems from introducing catastrophic errors or biases into national legal frameworks.

Body

Gen AI are not actually intelligent in the ordinary human sense and are unable to reliably answer questions that require a nuanced understanding of language content: par 1Gen AI chatbots predict the most likely combination of words, not necessarily the most correct or accurate answer.Limited training on Australian law and currencyGen AI responses may contain incorrect, opinionated, misleading or biased statements presented as factConfidentiality, suppression, privacy: par 2Ethical issues, including biases in training data, copyright and plagiarism, acknowledgment of sources: par 4