Photo: Shixart1985, CC BY 2.0, via Wikimedia Commons. Cropped from the original.
Many people assume that if an AI tool produces an image or paragraph, they can use it however they like. Others assume anything made with AI is legally off limits. Australian law sits somewhere in between, and parts of it are still unsettled.
The uncertainty is real but manageable. Some principles are already clear, including that Australia has no general fair use exception and that consumer law applies to misleading content however it was made. Other questions, such as who owns AI-generated work, are still being examined by government.
This guide explains where Australian copyright law currently stands on AI, what that means for using AI-generated images, video and text, and how consumer law applies to AI content and deepfakes. It is general information, not legal advice. For a broader introduction, see our guide to artificial intelligence.
How Australian copyright differs
Australian copyright law does not have a broad fair use defence like the one in the United States. Instead, it has narrower fair dealing exceptions for specific purposes such as research or study, criticism or review, parody or satire, and news reporting. That narrower framework matters for AI, because many overseas arguments about AI training rely on fair use.
Copyright also protects original works that come from human authorship. That principle shapes the debate about whether, and how, material generated by AI can be protected, which is one of the areas the government has said it will examine.
AI training and the government’s position
In 2025 the Productivity Commission floated a possible text and data mining exception that would have allowed AI developers to use copyright material for training without permission. On 26 October 2025, Attorney-General Michelle Rowland ruled out that exception, saying the government would not weaken copyright protections when it comes to AI.
The government instead convened its Copyright and AI Reference Group to look at three areas. These are whether a paid collective licensing framework or voluntary licensing should apply to AI, how to improve certainty about copyright in material generated with AI, and whether a small claims forum could make lower-value infringement cases cheaper to enforce. Those discussions were ongoing at the time of writing, so check for updates before relying on any position.
> Key quote: Australia has ruled out a text and data mining exception, but how copyright applies to AI-generated work is still being clarified.
Who owns AI-generated work
Ownership is the least settled question. Because Australian copyright depends on human authorship, purely machine-generated output may not attract copyright protection at all. Where a person makes significant creative choices, such as substantial editing, arranging or combining material, the position may be different, but the line has not been clearly drawn.
In practice, this has two consequences. Work you generate with AI may be harder to protect from copying by others. And the terms of the tool you used matter, because providers set their own rules about how outputs can be used. Read those terms, particularly for commercial use, and keep records of how important assets were created.
Using AI images and video commercially
Businesses can and do use AI-generated images and video, but a few precautions reduce risk. Avoid prompts that ask for a specific artist’s style, a real person’s likeness, a trademark or a well-known character, because outputs close to existing works or identities are where disputes are most likely. Check results for logos, watermarks or recognisable people before publishing.
Context matters too. AI images work well for concept art, backgrounds and illustrations. They are riskier for anything that implies a real product, place or result. We cover tool choices in AI image generators and AI video generators, and the advertising angle in AI ad creative.
Consumer law, disclosure and deepfakes
Copyright is only half of the legal picture. Government guidance on AI adoption notes that the Australian Consumer Law’s prohibitions on misleading and deceptive conduct can apply to AI outputs, including deceptive use of deepfakes and misleading silence about when AI is being used. An AI-generated before-and-after image, a fake testimonial or a product photo that shows features the product does not have could all create liability.
There is no single rule requiring every AI-assisted image to be labelled. The safer test is whether a reasonable customer would be misled without disclosure. Where AI changes what a customer believes about a product, a person or a result, disclose it or do not use it. Deepfakes also appear in scams that impersonate well-known people or businesses, so it is worth monitoring for misuse of your own brand and reporting it through Scamwatch.
A practical checklist before publishing
Before publishing AI-assisted images, text or video, a short checklist helps manage risk. First, read the terms of the tool you used, including who owns outputs, whether commercial use is allowed and whether the provider offers any indemnity. Terms differ between free and paid plans, so check the plan you actually use.
Second, check the output itself. Does it closely resemble a known artwork, character, logo or photograph? Does it include a real person’s likeness or voice? Could anyone reasonably think it shows something that did not happen? Any yes answer is a reason to pause, change the output or get advice.
Third, keep records. Note the tool, the date, the prompt and any human editing. If your business ever needs to show how a piece of content was created, or argue that a person contributed creative effort, these notes help. Finally, make sure claims in AI-assisted marketing are accurate, because consumer law applies regardless of how the content was made.
Common mistakes
The most common mistake is treating AI output as automatically free to use. Tool terms, third-party rights and consumer law all still apply. Another is assuming you own AI output in the same way as human-created work, which may not be the case.
Some businesses also use AI to create realistic people, places or results in advertising without considering whether customers would be misled. Finally, many teams keep no record of how assets were made, which makes it harder to answer questions later.
Where to start
Australian copyright law has not been rewritten for AI, and the government has ruled out a broad training exception while it works on clearer rules. Read the terms of the tools you use, avoid outputs that copy identifiable works or people, keep records and apply the consumer law test before publishing. For marketing help, contact us or email hello@thedigitalmarketingconsultants.com.au. This article is general information, not legal advice.
