The Use of Generative Artificial Intelligence in Family Law Matters: Risks, Limits and Court Expectations
Artificial intelligence is no longer a future concept in legal practice. Generative AI tools—such as ChatGPT, Microsoft Copilot and similar platforms—are already being used by lawyers to improve efficiency, reduce costs and manage increasingly complex legal work.
In family law, however, the use of generative AI requires particular caution. The confidential and sensitive nature of family law proceedings, combined with strict professional and court-imposed obligations, means that AI must be used responsibly, transparently and within well-defined boundaries.
This article examines how generative AI can be used appropriately in family law matters, and outlines the key guidance issued by Australian Courts.
What Is Generative AI?
Generative AI refers to artificial intelligence systems capable of creating new content—such as text, summaries or images—based on patterns learned from large datasets. In legal practice, generative AI can assist with tasks such as summarising material, drafting non-evidentiary documents, proof-reading and administrative automation.
While useful, generative AI does not think, reason or exercise legal judgment. It may produce inaccurate or entirely fictitious information if not carefully supervised.
For family law practitioners, AI should be used very carefully and only ever used as a support tool, not a decision-maker or substitute for legal expertise.
The Federal Circuit Court and Family Court of Australia (FCFCOA) Practice Direction on Use of Artificial Intelligence
The use of generative AI is now directly governed by the Federal Circuit and Family Court of Australia Practice Direction – Use of Artificial Intelligence which sets clear expectations for lawyers, litigants and experts appearing before the Court. The Practice Direction establishes a principled framework to ensure that any use of AI maintains the integrity of the justice system.
Other Courts such as the Federal Court of Australia and the Supreme Court of NSW have already issued Practice Notes to provide guidance on how AI may be used in those court proceedings.
All these Courts have acknowledged that generative AI can:
increase efficiency;
reduce costs; and
improve access to justice.
However, it emphasises that these benefits must never come at the expense of accuracy, fairness or public confidence in the justice system.
Core Principles for Users of AI in Court proceedings
The key expectations include:
1. Understanding Capabilities and Risks
If you use generative AI, you must understand its limitations, including its tendency to “hallucinate”—that is, generate plausible but false information, including fabricated case law or incorrect legal principles.
2. Responsibility for All Material Filed
Every user, including legal practitioners, are accountable and remains fully responsible for every document filed or submitted to the Court, regardless of whether AI was used in its preparation. AI use does not dilute professional obligations of diligence, competence or honesty.
Recent cases have shown serious consequences where AI-generated errors were not independently verified, including referrals to professional regulators.
3. Disclosure Where Required
Disclosure of AI use is generally required where generative AI has contributed to:
affidavits;
witness statements;
expert reports; or
other evidentiary material.
Where disclosure is required, it must clearly state how and where AI was used, ensuring the Court and other parties are not misled. The Supreme Court of NSW has prohibited using generative AI to prepare affidavit and witness statements.
4. Duties extend to Experts
Experts are also expected to comply with the FCFCOA Practice Direction in preparing reports and lawyers are required to provide a copy of the Practice Direction to any expert engaged.
Family Law–Specific Risks and Restrictions
Confidentiality is critical in family law, and many AI tools should therefore not be used in family law matters.
Open or public AI platforms may store or reuse data, which can create serious risks—particularly where court documents, disclosure material or subpoenaed records are involved. The conversations undertaken in open or public AI platforms can be subpoenaed and therefore extreme caution should be used when using open or public AI platforms. For example, if you upload your legal advice onto these platforms, you may waive your legal professional privilege.
Australian Family law proceedings are also subject to strict non-publication rules under the Family Law Act 1975 (Cth). Uploading Court material, subpoenaed documents or client disclosure material onto an open AI platform may constitute an unlawful communication to the public.
The FCFCOA has expressly warned against inputting confidential, privileged or suppressed material into AI systems where data storage and reuse cannot be controlled.
Appropriate Uses of Generative AI in Family Law
Used correctly, generative AI may provide meaningful assistance in non-evidentiary and administrative tasks, including:
creating internal chronologies based on existing documents;
drafting file notes or consultation summaries;
proof-reading correspondence;
diarising orders and key dates; and
assisting with document formatting.
Closed-system tools—such as enterprise-grade platforms integrated within secure legal software—offer stronger data protection and risk management than publicly accessible AI platforms.
What AI Should Not Be Used For
Consistent with court guidance and professional standards, generative AI should not be used for:
drafting affidavits or witness statements;
altering or “improving” a client’s evidence;
legal research without independent verification;
preparing legal advice; or
handling subpoenaed or confidential material in open systems.
AI-generated outputs must always be carefully reviewed by the responsible solicitor.
Looking Ahead
Generative AI will continue to evolve and will play an increasing role in legal practice. Courts have made it clear that they are not opposed to its use—but they expect lawyers and all users to adopt it thoughtfully, responsibly and in strict compliance with existing obligations.
In family law matters, where trust, confidentiality and accuracy are paramount, the careful and transparent use of AI is not optional—it is essential.
If you require advice about your family law matter or wish to understand how technology is used responsibly in your case, contact our experienced family law practitioners for tailored guidance.
Frequently Asked Questions About AI in Family Law
-
Yes—but only in a responsible and limited way. Family lawyers may use generative AI to assist with administrative and non‑evidentiary tasks such as summarising documents, preparing internal chronologies, proof‑reading correspondence and managing timelines.
AI cannot replace legal judgment and must not be used to give legal advice, draft affidavits or create evidence. Ultimately, your lawyer remains fully responsible for every document produced and every decision made in your matter.
-
Confidentiality is critical in family law, and not all AI tools are appropriate. Open or public AI platforms may store or reuse data, which can create serious risks—particularly where court documents, disclosure material or subpoenaed records are involved.
Responsible family law firms use secure, closed‑system AI tools and apply strict internal policies to ensure compliance with:
non‑publication provisions under the Family Law Act;
the Harman undertaking relating to subpoenaed documents; and
professional confidentiality obligations.
If AI is used at all, it should never compromise client privacy.
-
Courts acknowledge that AI can improve efficiency and reduce legal costs—but they also make it clear that accuracy and integrity are paramount. Generative AI must not be used to create or alter evidence, and lawyers remain personally accountable for ensuring all material filed with the Court is correct.
In some circumstances, a lawyer may be required to disclose the use of AI, particularly where it has assisted with evidentiary material. Failure to properly check AI‑generated content has already led to judicial criticism and disciplinary referrals.
