Skip to main content

Clifford Chance
Class Actions Insights<br />

Class Actions Insights

Australia’s Federal Court puts AI to the test

Two Australian class actions could mark a new phase in AI adoption by the judiciary, as the Federal Court of Australia explores technology's role in case management of complex litigation.

As businesses worldwide grapple with how to deploy AI responsibly, courts are beginning to ask their own version of the same question. To date, their attention has largely focused on regulating its use by lawyers and guarding against inaccurate submissions, fabricated authorities and undisclosed reliance on automated tools.

Two recent decisions of the Federal Court of Australia (a day apart) are of note. In Rogers v McDonald’s Australia Ltd (AI-use) [2026] FCA 1264 and R&B Investments Pty Ltd (Trustee) v Blue Sky Alternative Investments Limited (in liq) (Security for Costs and AI) [2026] FCA 1265, the Court directed parties in complex class actions to investigate whether AI could assist in preparing their cases for mediation and trial.

These decisions are pointing in a new direction. The Court has moved beyond setting boundaries around AI use. It is now asking whether AI can form part of the architecture of complex litigation. The approach adopted in Australia may provide a model for courts elsewhere as they confront similar questions of cost, complexity and scale.

From regulating AI to testing its value

In April 2026, the Court published its Use of Generative Artificial Intelligence Practice Note. The Practice Note recognised AI’s potential to improve efficiency and reduce legal costs, while establishing safeguards for its responsible use in proceedings.

In two class action matters, the Court has now made orders using its case management powers to require the parties to confer with the assistance of a suitably qualified technology expert selected by them. Together, the parties must examine whether AI could assist with:

  • collating and analysing data;
  • preparing chronologies and issue-based summaries of primary documents;
  • identifying areas of factual agreement and disagreement; and
  • managing and interpreting broader documentary records.

The parties must then report to the Court on the outcome of that process and/or provide a draft protocol.

A new model for litigation design

The decisions appear to be an evolution from regulating lawyers’ incidental use of generative AI to testing, and indeed encouraging, its use through court-supervised case management design.

The Court justified its departure from conventional practice by reference to the demands of modern litigation. The observation was that large legal teams can spend substantial time collecting, reviewing and synthesising documents before the issues that matter come into focus. The Court questioned whether these familiar but labour-intensive processes should remain the default where technology may perform some of that work more efficiently.

The Court also recognised that efficiency cannot come at the expense of reliability or fairness. It stated that any AI-assisted process must preserve traceability to primary records, permit scrutiny and correction, protect privileged and confidential material, and remain subject to human supervision. The Court was clear AI must not displace professional or judicial judgment.

This framework may find support beyond Australia. Expert-assisted conferral and a court-supervised protocol is an existing tool for Courts in areas such as discovery, although this model may provide other courts with a practical way to incorporate AI in complex litigation with the objective of achieving efficiency and without compromising procedural fairness. We expect too it will require adaptation to each jurisdiction’s procedural requirements.

What next?

The Federal Court has opened a door that courts elsewhere may not ignore. For businesses and legal teams, the challenge will be to identify where AI can make a real difference and to design an approach suited to the dispute. That will require early engagement between clients, lawyers and experts who understand both complex litigation and the practical, technical and governance issues that AI presents.

Note: Clifford Chance acts for the Fourth Respondent, EY, in the Blue Sky proceeding.

  • Share on Twitter
  • Share on LinkedIn
  • Share via email
Back to top