This comprehensive guide explains the current legal position in Australia, the specific risks each AI platform poses, and what you must never do when seeking legal advice or managing a dispute.

We reference four recent Australian court decisions, analyse platform data handling practices, and provide actionable guidance on protecting your confidentiality. Whether you are a business using AI, a lawyer advising clients, or a party involved in litigation, this article is essential reading.

An increasing number of clients and lawyers are uploading legal correspondence, advice, and confidential documents into artificial intelligence platforms such as ChatGPT, Microsoft Copilot, Google Gemini, and Claude. Many assume this is harmless. Many are wrong.

The consequences can be severe: the loss of legal professional privilege, breach of confidentiality obligations, and exposure in litigation discovery. This article explains the current legal position in Australia, the risks specific to different AI platforms, and what you must not do when seeking legal advice or managing a dispute.

The Fundamental Problem: How Legal Professional Privilege Works

Legal professional privilege is one of the oldest and most important protections in the Australian legal system. It protects confidential communications between you and your lawyer from being forced out in court or handed over to your opponents. The protection is not there to keep secrets: it exists so that you can seek frank legal advice without fear that your lawyer's views, your instructions, or the legal analysis will be used against you.

The privilege attaches to two categories of communication: advice privilege (communications for the dominant purpose of obtaining legal advice) and litigation privilege (communications made for the dominant purpose of actual or anticipated litigation). The High Court has confirmed that privilege is a substantive immunity against powers to compel disclosure, and it sits at the heart of access to justice.

But privilege is fragile. It depends entirely on confidentiality. Once you disclose a privileged communication to someone outside the lawyer-client relationship, privilege is lost. The question is not what you intended; the question is whether your conduct is inconsistent with keeping the communication confidential.

This is where AI becomes dangerous.

The Australian Legal Position: Recent Court Guidance

Until recently, Australian courts had not squarely addressed whether uploading legal documents to AI constitutes a waiver of privilege. That has changed. In 2025, Australian courts began issuing clear warnings.

Helmold v Mariya (No 2) [2025] FedCFamC1A 163

In this Federal Circuit and Family Court decision, the Full Court considered deficiencies in materials filed by an appellant, which the Court attributed to the use of generative AI. The appellant's materials contained incorrect and untraceable authorities; the Court concluded these errors could only have come from AI use.

More importantly, the Court commented directly on the privilege risks. The Court observed that the 'input of documents arising out of the proceedings into a generative AI program which stores, collates and replicates data may waive privilege or fall foul of the requirements that certain matters be treated as commercial in confidence'. This is the clearest statement yet from an Australian appellate court on the point.

The Court emphasised that these risks warrant care, and it endorsed recent judicial warnings about AI use. The court later referred the practitioners' conduct to the relevant professional regulatory bodies for investigation.

Why AI Platforms Pose a Unique Risk to Privilege

The Australian courts' concerns are grounded in how AI platforms actually operate. Consumer-grade AI systems are not confidential systems. Here is what happens to your data when you upload a legal document or email from your lawyer to ChatGPT, Gemini, Claude (free tier), or Copilot:

Consumer AI Platforms: The Default Position

ChatGPT (Free, Plus, and Team tiers): By default, OpenAI retains your conversations indefinitely unless you delete them. Even after deletion, OpenAI keeps the data on its servers for up to 30 days to detect abuse. More importantly, unless you have manually opted out in your settings, your conversations may be used to train OpenAI's models. This means your legal advice, your lawyer's analysis, your confidential instructions, and your business strategy become part of the training data for future versions of ChatGPT.

Claude (Free and Pro tiers): Anthropic, the maker of Claude, updated its privacy policy in late 2025. If you use Claude Free, Pro, or Max and have not disabled "Help improve Claude" in your settings, Anthropic retains your conversations for up to five years and uses them for training. Five years. Opt out, and retention drops to 30 days.

Google Gemini: Google's policy varies depending on which product you use. The free Gemini Chat defaults to training on your data; Google Workspace Gemini (for business users) does not.

What You Should Never Do

Do not upload copies of legal advice we have given you to any AI platform,

Do not upload your correspondence with us to ChatGPT or similar services, even to ask for a summary or a second opinion.

Do not use AI to draft responses to our advice without considering that the AI may store your material.

Do not share draft contracts, settlement proposals, or litigation strategy with AI platforms.

Do not ask an AI to review or redraft legal documents we have prepared, even if you think you are paraphrasing.

How Oceania Legal Can Help

If you are uncertain about your use of AI or the confidentiality of your legal communications, we are here to advise. We routinely advise clients on data protection, confidentiality obligations, and risk management in commercial dealings.

We offer advice on designing and implementing AI policies that protect legal privilege and client confidentiality; review and drafting of data processing agreements with AI providers; risk assessment for organisations considering AI adoption; and counsel on disclosure obligations and remediation if confidential information has been inadvertently disclosed.

Get in touch with the team at Oceania Legal to arrange a confidential discussion. We are commercial lawyers who understand the intersection of technology, risk, and business. We can help you navigate this landscape with clarity and confidence.

Visit our services page to learn more about how we advise clients on commercial risk and compliance, or contact us directly to discuss your needs.