This article summarises publicly available guidance from regulators and official sources. It is general educational information only and does not constitute legal or professional advice. Requirements vary by jurisdiction. Consult your regional authority or a qualified professional for advice specific to your situation.
AI language models hallucinate: they generate confident, well-written, and completely fabricated information. For most personal AI use, this is an inconvenience. For a business that publishes AI-generated content about its products, services, prices, or competitors, a hallucination in the published content can become a misleading representation under the Australian Consumer Law. The law does not require the business to have intended to mislead anyone. If the conduct is likely to mislead or deceive a reasonable person, the ACL concern exists regardless of how the false content was produced.
In short: The Australian Consumer Law prohibits misleading or deceptive conduct in trade or commerce, and false representations about goods or services. These provisions apply to content a business publishes regardless of whether AI or a human wrote it. AI hallucinations create practical risk when published content contains wrong prices, wrong product features, wrong regulatory information, wrong statistics, or false claims about competitors. The practical response is a verification step before publication, not an avoidance of AI tools altogether. This guide explains the regulatory framework and what a reasonable verification approach looks like.
Last reviewed: June 2026 | Next review: December 2026
What AI Hallucinations Actually Are
AI language models generate text by predicting the most likely next word based on patterns in their training data. They do not retrieve facts from a verified database. This means they can produce content that is factually wrong but linguistically plausible: a statistic with a plausible-sounding source that does not exist, a product feature that is slightly wrong, a regulatory requirement that is misattributed, or a competitor claim that is simply made up. The term 'hallucination' refers to this class of output: confident-sounding text that does not reflect reality.
Hallucinations are more common in some content types than others. Legal and regulatory information is particularly prone to hallucination because AI models have seen large volumes of legal text but cannot always distinguish which specific provisions apply to which specific situations. Statistics and research citations are prone to hallucination because AI models generate citation-shaped text rather than retrieving actual research. Product pricing is prone to hallucination because prices change and AI training data goes stale. Competitor claims are prone to hallucination because AI models are generating what sounds like competitive analysis rather than checking current facts.
These categories matter for the ACL risk assessment because they correspond directly to the types of representations that the ACL is most concerned with: pricing, product features, comparative claims, and regulatory compliance. If your business uses AI to draft content in these categories, the hallucination risk is highest in exactly the areas where the ACL liability risk is also highest.
What the Australian Consumer Law Says
The Australian Consumer Law is Schedule 2 to the Competition and Consumer Act 2010, and it is enforced primarily by the ACCC (Australian Competition and Consumer Commission). Two provisions are most relevant to AI-generated content.
Section 18: Misleading or deceptive conduct. The ACL states that a person must not, in trade or commerce, engage in conduct that is misleading or deceptive or is likely to mislead or deceive. The ACCC's guidance notes that this is an objective test: what matters is whether a reasonable person in the audience would be misled, not whether the business intended to mislead. Publishing factually wrong information about your products, services, prices, or the nature of an offer can constitute misleading or deceptive conduct if it is likely to mislead a reasonable person, even if no one is actually misled in a given instance. ACCC guidance on misleading conduct is at accc.gov.au/business/anti-competitive-behaviour/false-and-misleading-conduct.
Section 29: False or misleading representations. The ACL also specifically prohibits false or misleading representations about goods or services, including representations about price, quality, standard, grade, value, composition, performance characteristics, and country of origin. Comparative advertising claims that falsely represent a competitor's products are also covered. If AI-generated content makes specific false representations about any of these things, section 29 is also potentially engaged.
Neither provision requires intent to mislead. Neither provision is avoided by explaining that AI wrote the content. The business that publishes the content is the conduct entity under the ACL, and the business bears responsibility for what it publishes.
Where AI Hallucinations Create ACL Risk
Product and service descriptions. Using AI to draft product listings, service descriptions, capability claims, or feature lists is common and generally appropriate. The risk is that AI may generate feature claims that are slightly or significantly wrong. A product listing that says a device has a feature it does not have, or that a service includes something it does not, is a straightforward misleading representation. Human review of all specific product or service claims before publication is the relevant safeguard.
Pricing information. AI-generated content about pricing is particularly prone to hallucination because pricing data goes stale quickly. If a business uses AI to draft a comparison of competitor prices and the AI generates wrong figures, the resulting publication could be a false representation about a competitor. Pricing claims in any AI-generated content should be verified against current sources before publication.
Regulatory and compliance claims. Many businesses use AI to generate content about their compliance status, certifications, standards they meet, or regulatory requirements their products satisfy. AI can confidently generate wrong regulatory information. A claim that a product meets a standard it does not meet, or that a service is compliant with a regulatory requirement it has not actually been assessed against, is a potentially serious ACL issue. Regulatory claims should be verified against the relevant primary source before any publication.
Statistics and research claims. AI tools frequently hallucinate statistics, including generating plausible-looking figures with invented sources. Publishing a statistic from a source that does not exist, or attributing a figure to a study that does not say what the AI claims, can constitute misleading conduct if the context creates a false impression of authority. Any statistics or research references in AI-generated content should be verified against the original source.
Comparative advertising. Using AI to draft comparative content about competitors is a higher-risk category, because the ACL specifically covers false or misleading representations about competitors. If AI generates a comparison that misrepresents a competitor's pricing, features, or capabilities, that content engages both section 18 and potentially section 29. Comparative claims about named competitors require factual verification before any publication.
What This Looks Like in Practice
Consider a small business that uses AI to draft content for its website, including a page comparing its product to two competitors. The AI generates the page with specific pricing and feature comparisons. Some of the competitor information is out of date or wrong. The business publishes without checking. A competitor sees the page, identifies the false claims, and contacts the ACCC. The business faces the potential cost of compliance correspondence, required corrections, and depending on the severity and reach of the content, potential enforcement action.
Or consider a healthcare practice that uses AI to draft a web page about a treatment service. The AI generates a claim about clinical effectiveness that is not supported by current evidence, or cites a statistic from a source that does not exist. A patient reads the page and makes a decision based on that content. The ACL and potentially AHPRA's advertising rules are both engaged.
These are not hypothetical edge cases. They reflect normal patterns of AI-assisted content production applied to common business publishing activities. The legal framework that applies is not new: the ACL has applied to misleading business conduct since 2011. What is new is that AI tools create a much faster route to publishing misleading content at scale, without the natural friction that usually occurs when a human writes and reviews content before publication.
What Businesses Can Do
The practical response to AI hallucination risk under the ACL is not to avoid AI tools for content. It is to build a verification step into the publishing workflow for content that makes factual claims. This does not require extensive legal expertise. It requires the person reviewing the content to ask a specific question before publication: is every factual claim in this content something I have actually verified against a primary source?
Claims that require verification before publishing AI-generated content include: prices (current, verified against actual pricing), product specifications and features (verified against the actual product or vendor source), regulatory status (verified against the relevant regulator or certification body), statistics (verified by checking the original source document), and competitor claims (verified against the competitor's own current information). Where these can be checked, check them. Where they cannot be checked or the source cannot be located, remove the claim.
AI tools are genuinely useful for drafting structure, tone, and non-specific content. They are not reliable for generating factual claims without verification. A practical AI content workflow treats AI-generated text as a draft that needs a facts pass, not as a finished product ready to publish. Our AI content verification checklist provides a practical step-by-step format for this review process.
What the ACCC Has Said About AI and Business Conduct
The ACCC has noted in its Digital Platform Services Inquiry and other public communications that it is monitoring AI-related consumer harm, including the potential for AI-generated content to constitute misleading conduct under the ACL. As of June 2026, the ACCC has not published specific guidance on AI-generated content and misleading conduct, but the existing ACL framework applies fully. The ACCC has indicated that the use of AI in content production does not create a special exemption from existing consumer law obligations.
The ACCC's online resources for businesses on misleading conduct are at accc.gov.au/business. The ACCC's Digital Platforms Branch publishes regular updates on digital economy enforcement priorities, which are worth monitoring if your business uses AI tools for significant customer-facing content production. For professional services businesses, your relevant industry regulator may also have specific guidance on AI content.
Methodology (Real-World, Verified)
We test AI tools against real SMB workflows: the tasks a 20-person business actually uses AI for, not enterprise demos. Pricing is verified in AUD at the vendor's published rates or converted at current exchange rates. Compliance notes reference the legislation and regulatory guidance relevant to each article's scope. Tools are assessed for suitability by a business with no dedicated IT department.
Can a business be liable under the ACL if AI generated the wrong content?
Yes. The ACL applies to conduct in trade or commerce, and a business that publishes false or misleading content is responsible for that conduct regardless of whether a human or an AI tool produced the content. The ACL's misleading conduct provision (section 18) is an objective test based on whether a reasonable person would be misled, not on the intent or process of the business. The practical implication is that AI-generated content requires the same fact-checking discipline as human-written content before publication, particularly for specific claims about prices, features, regulatory status, and competitors.
What is an AI hallucination?
An AI hallucination is when an AI language model generates text that is factually wrong but linguistically confident. Language models generate text by predicting the most likely next word based on training data patterns, not by retrieving verified facts from a database. This means they can produce wrong statistics, invented source citations, incorrect product details, and false claims about regulatory requirements, all written in the same fluent and confident style as accurate content. Hallucinations are more common with specific factual claims (prices, statistics, regulatory details, competitor information) than with general structural or stylistic content.
Does the ACL apply to content I publish on social media?
Yes. The ACL applies to conduct in trade or commerce, and content published by a business on social media in a commercial context is conduct in trade or commerce. A false product claim in an Instagram post, a wrong price in a Facebook advertisement, or misleading comparative claims in a LinkedIn post are all covered by the ACL's misleading conduct provisions. AI-generated social media content requires the same factual verification discipline as website content, particularly for specific product claims, pricing, and comparative statements.
Is there a legal requirement to disclose that content was written by AI?
As of June 2026, there is no specific Australian law requiring businesses to disclose that content was produced with AI assistance. However, if AI-generated content creates a false impression that it was written by a human expert in circumstances where that matters to a reader (for example, a medical article attributed to a named doctor that was generated by AI), this could engage the misleading conduct provisions by virtue of the false impression created. The ACCC has not published specific guidance on AI disclosure obligations, but this is a developing area. For any content where the method of production is likely to be material to how a reader interprets and relies on it, disclosure is a reasonable approach.
What is the difference between AI hallucination risk and copyright risk for AI content?
They are different types of risk. Hallucination risk is about factual accuracy: AI generates false information that could mislead customers and create ACL liability. Copyright risk is about reproduction: AI may generate content that closely reproduces someone else's copyright-protected material, creating potential infringement liability. Both risks exist in AI-generated content, but they are managed differently. Factual verification addresses hallucination risk. Copyright considerations are relevant to any content that closely resembles existing published material, and the position on AI-generated content and copyright in Australia is still developing. Our guide on AI-generated content and legal liability in Australia covers both dimensions.
The information in this article is general in nature. It reflects a summary of publicly available guidance and does not constitute legal, privacy, or professional advice. Your obligations will depend on your specific situation, jurisdiction, and business circumstances. Do not rely on this article as a substitute for qualified legal or professional advice.
A practical verification workflow before publishing AI-generated content is the most effective response to ACL hallucination risk. Our AI content verification checklist provides a step-by-step format for checking factual claims before publication.
Get the AI Content Verification Checklist