AI caller disclosure in Australia: must you say it's AI?
Published 6 October 2026 · Sources read on 5 and 6 October 2026 · By Target Digital · 8 minute read
AI caller disclosure is not required by any Australian law we could find, but consumer law makes it the safe choice, and the telemarketing rules already assume a synthetic caller will identify the business and the purpose of the call. The practical answer is to say it is an AI in your first sentence.
This guide sets out what the sources say, where the gaps are, and the words to use on the call. It is general information, not legal advice. The wider rules on consent, hours and recording are in our AI outbound calling compliance guide.
AI caller disclosure: what the law says
We found no Australian statute that requires an AI caller to announce that it is an AI. The government's National AI Plan says its regulatory approach will continue to build on existing legal and regulatory frameworks, which remain the foundation for addressing AI-related risks.
Treasury reached the same place on consumer law. Its Review of AI and the Australian Consumer Law, released in October 2025, found that the Act's prohibitions on misleading or deceptive conduct apply equally to AI-enabled goods and services. The minister's release said Australians have the same consumer protections for AI as for traditional goods and services, and put it as consumer laws being "strong enough to cover new technology, with only minor tweaks identified".
So the question is not whether a rule says "disclose". It is whether staying silent, or answering a direct question falsely, breaks a rule that already exists.
How misleading conduct applies to an AI caller
The Australian Consumer Law prohibits misleading or deceptive conduct in trade or commerce. The ACCC notes that "It makes no difference whether a business intends to mislead or not." The Treasury review adds three points that matter here:
- It applies to AI. The prohibitions on misleading or deceptive conduct and false or misleading representations apply equally to AI-enabled goods and services.
- Silence can mislead. The review explains that under the Australian Consumer Law, silence or omission may be misleading "where there is a reasonable expectation that a fact, if it exists, will be disclosed". That expectation is judged objectively, case by case, and the law does not require disclosure with total disregard for the business's own interests.
- The business carries the risk. A person can breach the rule without fault, and the review says the onus is on businesses to make sure the technologies they use are fit for purpose.
Put those together. Whether a caller expects to be talking to a person is a question of fact. If your agent uses a human name, a casual manner and no introduction as an AI, a lead could reasonably assume a person. If the lead then asks "Am I speaking to a real person?" and the agent says yes, that is a false statement, and no review is needed to see the problem.
What the telemarketing rules already require
The Telemarketing and Research Calls Industry Standard 2017 sets what a telemarketing caller must say as soon as the call starts. Three details apply to AI:
- The caller gives the business name, or a name by which it can be readily identified, and the purpose of the call. That is a requirement whether the voice is human or synthetic.
- The caller's own given name is required, unless the call is made solely using a recorded or synthetic voice. The rule contemplates synthetic callers and does not ask them for a human first name.
- For calls with a recorded or synthetic voice, section 12 requires a mechanism during the call to request the contact and complaint details, such as pressing a button to talk to an operator.
Two readings follow. The minimum the Standard asks for can be met with "this is the assistant for [business], calling about your enquiry". And because a synthetic caller is not asked for a human name, inventing one adds risk and no legal benefit. That second point is our view, not a rule.
What government guidance says
The government's Guidance for AI Adoption says "People should know when they're interacting with AI" and tells organisations to make it standard practice to communicate their use of AI to stakeholders. It is guidance, not law. It does tell you where regulators and customers are heading.
The privacy regulator says the same for businesses that handle personal information. The OAIC's guidance on commercially available AI products says businesses should update their privacy policies and notifications with clear information about their use of AI, including making sure public-facing AI tools such as chatbots are clearly identified as such to external users such as customers.
One distinction to keep clear: in 2026 ACMA registered a commercial radio code that requires stations to disclose a synthetic voice hosting a program. That is a broadcasting rule and does not cover phone calls.
What to say on the call
These are starting points to adapt with your adviser.
- Opening line: "Hi, this is the AI assistant for Smith Plumbing, calling about the quote you asked for. This call is recorded. Is now a good time?"
- If asked "Are you a real person?": "No, I'm an AI assistant. I can book you in, or I can connect you to a person on the team now."
- If asked for a human: offer a live transfer or a call back from a named person with a time. Do not argue.
- If they say stop or "not interested": end the call straight away and suppress the number. The Standard requires the call to end immediately when the person asks or shows they do not want to continue.
Avoid a human first name for the agent, claims of being a staff member, and any scripted answer to "are you a robot?" that dodges the question.
Put it in your consent wording too
The disclosure should not be the first the lead hears. Add it to the form: "Yes, [your business] may call and text me about this enquiry, including using an AI assistant." Our guide on consent to call from a Facebook lead form explains where it goes.
Check your provider's terms
Read who carries the duty. AiDial's published FAQ (read 5 October 2026) says the customer is responsible for informing callers they are interacting with an AI agent, where required by law, and for ensuring appropriate consent is obtained where required. Other providers may say the same. Ask each one in writing what its agent says by default, whether you can approve the script, and whether the disclosure can be switched off.
Whichever provider you use, including us, ask to approve the opening script before it goes live. Our AI Outbound Call Engine exists to call fast, and the case for that is in our speed to lead guide. Fast is only useful when the call is honest.
AI caller disclosure: common questions
Does an AI caller have to say it's an AI?
We found no law that says so outright. The consumer law ban on misleading conduct, the telemarketing Standard and government guidance all point towards telling people, and a false answer to a direct question is the clearest risk.
Can an AI agent pretend to be a person?
We would not do it. Treasury's review says misleading conduct can occur without intent and that the business carries the onus. An agent that claims to be human when asked is making a false statement.
Does the Standard require the AI to give a name?
It requires the business name and the purpose of the call. It does not require a given name when the call is made solely with a recorded or synthetic voice.
Is this likely to change?
Possibly. The government says it is continuing to assess existing laws for AI, and the Federal Register lists the telemarketing Industry Standard as due to be repealed on 1 April 2027 under the sunsetting rules. Check ACMA and Treasury before you rely on this page after October 2026.
Next step
If your agent already calls leads, a 30-minute call is enough to review its opening script, its answer to "are you a robot?" and your form wording. We will tell you what we would change. This is general information, not legal advice.