AI Law - International Review of Artificial Intelligence LawCC BY-NC-SA Commercial Licence ISSN 3035-5451
G. Giappichelli Editore

21/11/2026 - AI-Assisted Court Claims Expose Self-Represented Litigants to Legal and Financial Risks (Australia)

argument: Notizie/News - AI in Judicial Activities

Source: UNSW Sydney

Felicity Bell, Michael Legg and Vicki McNamara argue that self-represented litigants’ misuse of generative AI can undermine claims and increase courts’ workload. The researchers’ global database includes 295 Australian cases through 30 September 2026 in which courts or tribunals identified problematic AI use in proceedings. Self-represented litigants account for 81% of the Australian misuse cases recorded in that database. Examples include fabricated authorities, genuine authorities irrelevant to the dispute, and inaccurate AI-assisted witness statements or submissions. The authors acknowledge that AI can assist with legal work, but stress the need to check sources, verify information and explain an argument’s relevance. In Sadnan Khan v Aldi Pty Ltd, the Fair Work Commission ordered a former Aldi employee to pay A$1,230 in costs. According to the account, the commission found that Khan relied on AI-generated arguments despite circumstances in which Khan should have recognised the claim could not succeed; leave to appeal was refused.

The authors interpret recent Federal Court data as showing that appeals from unsuccessful litigants account for a substantial share of increased filings. The authors warn that pursuing proceedings in higher courts can increase litigants’ financial exposure. Two unpublished papers examining US civil court data reportedly identify increased AI-assisted filings by self-represented litigants without an apparent improvement in their low success rates. Most Australian courts have issued guidance on responsible generative AI use; Fair Work Commission guidance effective from 20 October 2026 addresses genuine legal sources and chatbot prompting. The responses described include opportunities to correct documents, warnings, restrictions on further filings, striking out claims or defences, and costs orders. Bell, Legg and McNamara argue that warnings and sanctions alone are unlikely to resolve misuse, given chatbots’ availability and encouraging responses. The authors propose improving litigants’ knowledge and support so that AI assistance strengthens participation in the justice system.