Unlike the Federal Court, NSW, Queensland, Victoria and South Australia, the FCFCOA has issued no formal practice note regulating how practitioners use AI in proceedings. What it has published is an AI Transparency Statement (April 2025), covering the court's own internal use of the technology for administrative purposes.
That gap is relevant: the FCFCOA handles proceedings where the quality of evidence determines outcomes for children, and therefore where document reliability carries stakes that most litigation doesn't. It is one of the courts where the risks of unverified AI-generated content are highest, and currently one where practitioners are navigating those risks without a defined framework to fall back on.
Most areas of legal practice carry a professional liability dimension to getting AI wrong. Family law adds something on top of that. Parenting disputes turn on the quality of evidence judicial officers rely on to form views about a child's best interests: affidavits, expert reports, the reliability of what each party says happened. Property proceedings require detailed financial disclosure that, if AI-assisted, has to be accurate at the particulars. Errors in other practice areas can often be caught and corrected mid-process. In family law, by the time an error surfaces, it may already have shaped a decision about where a child lives or how an asset pool is divided.
Neither consideration is a reason to avoid AI in family law practice. Both are reasons to be precise about which tools are appropriate for which tasks.
A contested property matter can span years of financial records across multiple parties. Reading that volume manually and catching every inconsistency is difficult at scale. Legal AI platforms can work across that material systematically. This might involve flagging discrepancies, surfacing relevant entries, and processing what human reviewers run out of time to read carefully. The practitioner still determines what the inconsistencies mean and how to deploy them in proceedings, but the tool produces a structured first pass that makes that analysis possible in the first place.
The body of decided family law case law is substantial, and the relevant question in practice is rarely what the statute says. It's how the court has applied the statute over time. A practitioner advising on a property settlement needs to know how courts have treated comparable asset structures, contribution patterns, and future needs arguments in past decisions. The same applies in parenting matters, where the 'best interests' framework is well-established but its application depends heavily on analogous facts. Keyword search across case databases yields inconsistent returns for this kind of comparative work, and the volume makes manual survey impractical. A retrieval-grounded research tool can surface relevant authority across a full corpus and synthesise how courts have approached a particular issue — giving the practitioner a foundation to build on rather than a set of links to sort through.
Research feeds into strategy, and this is where AI can do more than retrieve. A practitioner preparing for a final hearing in a parenting matter needs to understand not just which cases are relevant, but how the court has weighed particular risk factors, what expert evidence has carried weight in comparable scenarios, and where the opponent's position is most vulnerable on the authorities. In property matters, the same applies to contribution-based arguments, the treatment of inheritance, post-separation contributions, and how the s 75(2) factors have been applied in analogous circumstances. AI that can synthesise across primary and secondary sources in real time can compress hours of preparation into a focused, citable brief.
Family law generates a high volume of procedural and substantive documents — affidavits, submissions, financial statements, consent orders. AI can assist with structuring first drafts, ensuring statutory requirements are addressed, and cross-referencing factual claims against source material. The value is less about generating prose and more about reducing the mechanical overhead of document production so practitioners can focus their time on the parts of drafting that require judgment: how to frame the narrative, what to emphasise, what to leave out.
Tools like Habeas aid greatly in this process, by allowing family lawyers to upload their own precedents to the platform and use them as templates for further drafting.
For a general-purpose large anguage model (LLM) not specifically grounded in Australian case law, a question about the 'best interests' principle will produce an answer that sounds confident and competent. It may reference actual cases, cite provisions accurately, and still include decisions that don't exist.
The problem is that AI hallucinations aren't always obviously wrong. A practitioner relying on these outputs without verification is operating on false certainty, in a setting where the cost of that approach is high.
A retrieval-grounded tool like Habeas works differently. It searches a specific dataset, retrieves actual documents, and grounds its outputs in what those documents say. Verification, or further deep reading, is encouraged at the base layer of the platform because you're checking against a cited source rather than reconstructing the research from scratch. Advanced legal AI tools like Habeas can also synthesise information across primary and secondary sources in real time, enhancing both the accuracy and the depth of the analysis produced.
The FCFCOA's position is not that AI has no place in family law practice. Its concern is that some of the tools lawyers are using cannot be easily verified, or can produce non-existent citations, and that concern is proportionate to what's actually at stake in its proceedings. The architecture underneath the platform your firm uses matters, and getting that decision wrong in this practice area carries consequences that compound quickly.
Habeas retrieves from a verified dataset of Australian legislation and case law, with citations traceable to the source. See how it works or book a demo for your practice today.
