Artificial intelligence (“AI”) is rapidly becoming a go-to tool for people trying to navigate legal disputes without the cost of a lawyer. But a recent decision of the Federal Circuit and Family Court of Australia demonstrates very clearly why that approach can actually backfire.
In Jenson & Lockridge (No 2) [2026] FedCFamC1A 44, the Court dismissed an appeal where a self-represented litigant relied heavily on AI to prepare her case. The judgment is one of the clearest judicial warnings to date about the risks of using artificial intelligence in litigation.
Case summary:
This matter involved an appeal from property orders made under the Family Law Act 1975 (Cth).
The appellant, a self-represented litigant, relied on a “professional” artificial intelligence system to formulate her grounds of appeal, draft her Summary of Argument, and assist with her oral submissions.
At first glance, this might seem like a practical solution. However, the Court found that the use of AI interfered with the proper conduct of the appeal and ultimately weakened the appellant’s case.
As a result, the appeal was dismissed in its entirety, the appellant’s application to give further evidence was rejected, and she was ordered to pay the respondent’s costs in the sum of $1,540.
Why did the appeal fail?
Appeals are not an opportunity to re-run a case. To succeed, a party must identify a specific legal error made by the trial judge, consistent with well-established principles such as those in House v The King.
The appellant did not do this. Instead, her AI-generated material created significant legal problems.
The Court identified that the AI:
- Attributed reasoning to the trial judge that did not exist in the judgment (commonly referred to as “hallucinations”);
- Combined unrelated legal principles without proper context;
- Raised arguments that were inconsistent with the case presented at trial; and
- Obscured the real issues, making the appeal difficult to understand and assess.
In practical terms, this meant the Court was not presented with a clear, legally coherent argument identifying error. As a result, the appeal failed.
This highlights a critical limitation of AI, it may generate arguments that sound persuasive, but it does not understand whether those arguments are actually legally available.
A hidden risk: confidentiality and criminal liability
Perhaps the most serious, and least understood, risk relates to privacy and confidentiality.
The Court noted that uploading documents into AI systems may raise concerns under section 114Q of the Family Law Act 1975 (Cth).
What is section 114Q?
Section 114Q makes it a criminal offence to communicate to the public an account of family law proceedings that identifies:
- A party;
- A witness;
- A child; and
- Or anyone connected to the case.
The maximum penalty is 12 months imprisonment.
Importantly, a person can be “identified” even without being named. Identification can occur through descriptions, relationships, employment details, locations or any combination of information that allows someone to work out who the person is.
Why does this matter when using AI?
Here’s where the risk becomes very real. When you input documents into an AI tool, you may be uploading affidavits, sharing court orders, describing family members or children or explaining sensitive allegations
Even if you don’t intend to “publish” anything, you are still communicating that information to an external system
Depending on the platform, this may:
- Store your data;
- Process it externally; and
- Expose it to third-party systems.
This creates a serious legal question:
Have you just communicated an identifiable account of proceedings to “the public”?
While the law in this area is still evolving, the Court has already flagged this as a real risk. AI platforms are not designed to ensure compliance with family law confidentiality obligations, this responsibility remains entirely with you.
The financial consequences
The appellant’s appeal was not only unsuccessful, but it also resulted in a costs order against her.
This serves as an important reminder for clients that attempting to save money by relying on artificial intelligence instead of obtaining proper legal advice can have the opposite effect.
Rather than reducing expense, it can lead to a failed case, significant delays and wasted time, increased legal costs, and ultimately an order requiring payment of the other party’s costs.
Why engaging a lawyer matters?
The decision highlights an important reality: legal proceedings, particularly in family law, are not simply about presenting arguments. They require careful judgment, strategy, and a deep understanding of both the law and the procedural framework.
A lawyer does far more than draft documents. They assess the strengths and weaknesses of your case, identify which arguments are legally available, and ensure that those arguments are consistent with the way your case was run at trial.
A lawyer also ensures compliance with strict court rules. These are not technicalities; they are essential to how courts determine cases. Even a strong argument can fail if it is not presented correctly.
Importantly, lawyers exercise professional judgment. They know when an argument is persuasive, when it is risky, and when it should not be made at all. Artificial intelligence cannot make these distinctions. It may generate content that appears convincing, but it cannot evaluate whether that content is accurate, relevant, or strategically sound.
Ultimately, engaging a lawyer is not simply about legal knowledge, it is about protecting your position. It ensures that your case is presented clearly, strategically, and in a way that gives you the best possible chance of success, while avoiding unnecessary risks, costs, and errors.
If you want expert legal advice from a leading family or criminal lawyer, please contact the team at Barker Evans on (02) 8379 1892 or info@barkerevans.com.au
By Alyssa Maschmedt