Can AI Make Litigation Cheaper? The Federal Court Thinks It’s Worth Asking

By Winterkorn Legal Group

For the past few years, most of the conversation about artificial intelligence in Australian courtrooms has been a cautionary one. Judges have criticised fake case citations, courts have issued practice notes, and practitioners have been reminded that they remain personally responsible for everything they file.

A decision handed down in August 2026 changes the tone of that conversation. In Rogers v McDonald's Australia Ltd (AI-use) [2026] FCA 1264, Justice Lee of the Federal Court did not ask whether AI had been misused. Instead, his Honour directed the parties to work out whether AI could help them run the case more efficiently.

For businesses and individuals involved in disputes, and particularly those facing document-heavy litigation, the decision is worth understanding.

The dispute

The proceeding is a class action brought by three individual applicants and the Shop, Distributive and Allied Employees Association against McDonald's Australia Ltd and a group of franchisee companies. It alleges that managers at corporate-owned and franchised restaurants were not paid for work performed before and after their rostered shifts.

To get the matter ready for mediation, and for a ten-week initial trial listed for early 2028, the parties need to assemble detailed information about each group member: their roles, award or enterprise agreement classifications, shifts and pay. Justice Lee described that material as "dispersed through a miscellany of business records, apparently maintained in different repositories and in different forms" which must be collected, matched to individuals and periods, analysed and synthesised before it becomes useful.

Done manually, that is an enormous and expensive exercise.

What the Court ordered

Using the broad case management power in section 37P(2) of the Federal Court of Australia Act 1976 (Cth), Justice Lee ordered the parties, by 6 November 2026, to confer with a suitably qualified technology expert of their choosing about whether AI is suitable for:

• collating and analysing quantification data for mediation;
• preparing chronologies and summaries of primary documents by reference to specified issues; and
• handling and interpreting documents and other data in the proceeding more generally.

The parties must then give the Court a report, a draft protocol, or both, having regard to the Federal Court's Generative AI Practice Note (GPN-AI) published in April 2026 (Proctor). The direction is widely regarded as the first of its kind under section 37P(2).

Importantly, the Court did not order the parties to use AI. Justice Lee said he was not "prescribing the technology, selecting a provider or determining in advance the form of any eventual protocol". What mattered was that AI "be examined seriously, cooperatively and with an eye to the overarching purpose, rather than dismissed because the conventional alternative is known".

The four safeguards

Justice Lee was clear that AI is not a substitute for judgment, and that no court should hand a machine any substantive adjudicative task. The aim is to use technology for work that does not require human judgment, so lawyers and judges can concentrate on the work that does.

Any AI-assisted process in the case must therefore:

1. Be traceable. Every output must be capable of being linked back to the primary records it came from.
2. Be testable. Outputs must be capable of being checked and corrected.
3. Protect confidentiality. Privileged and confidential material must be safeguarded.
4. Remain under human supervision. Lawyers stay responsible for the work.

These four principles are a practical checklist for anyone thinking about using AI in a dispute, whatever the size of the matter.

Not a one-off

Rogers is part of an emerging pattern. A day later, the Court made similar directions in another class action, R&B Investments Pty Ltd (Trustee) v Blue Sky Alternative Investments Limited (in liq) (Security for Costs and AI) [2026] FCA 1265.

More recently, Justice Cheeseman relied on Rogers in a dispute where one side had used AI to summarise more than 7,500 documents. Her Honour observed that if one party can properly use AI to organise a large body of documents, subject to verification, "there is no principled reason" the other side should not use equivalent technology to speed up its own review. In other words, courts may become less patient with parties who rely on slow, costly manual processes when a verified technological option is available.

What this means for you

If you are a business involved in, or anticipating, a dispute:

• Expect courts to ask about efficiency. In matters involving large volumes of emails, contracts, payroll or financial records, the question of how documents will be handled may become part of case management.
• Your records matter. The difficulty in Rogers arose from data spread across different systems and formats. Well-organised, accessible business records make any review, manual or AI-assisted, faster and cheaper.
• Ask your lawyers how they will manage the material. Appropriately supervised technology may reduce the time spent on mechanical tasks such as sorting documents and building chronologies, leaving more of your budget for strategy and advice.

If you are concerned about AI and your confidential information:

The confidentiality safeguard is not a formality. Before client material is put into any AI system, a law firm should understand where the data will be processed and stored, whether it will be retained or used to train the system, and who can access it. Those questions also bear directly on legal professional privilege.

A word of caution on reliability:

AI output can look complete and polished while quietly leaving things out. That is a different problem from a fabricated case citation, and it can be harder to spot. This is why traceability and testing matter: when every point in a chronology or summary links back to a source document, gaps are easier to find. Responsibility for the final work product always rests with the lawyer.

The bottom line

Rogers does not mean AI must be used in every case, and it does not replace the professional obligations that apply to lawyers. But it signals a real shift. In document-heavy disputes, the Federal Court now expects parties to at least consider whether technology can do the mechanical work more quickly and cheaply, provided the results are traceable, testable, confidential and properly supervised.

For clients, that is good news: the potential for lower costs and faster resolution, without compromising the quality of legal judgment.

How Winterkorn Legal Group can help

Winterkorn Legal Group acts for businesses and individuals in commercial and civil litigation across Queensland and Australia. We take a practical, commercial and pragmatic approach to disputes.

If you are involved in a dispute or would like advice on your options, contact our team to arrange a consultation.

This article is general information only and does not constitute legal advice. You should obtain advice specific to your circumstances. Information is current as at October 2026.