Do you have to say a call is an AI in Australia?

Is there an Australian law that makes you say it is an AI?

Not one we could find, and we went looking in the places it would be.

Australia regulates AI calling through laws written before AI calling existed. The Do Not Call Register Act 2006 governs which numbers may be called, the Telecommunications (Telemarketing and Research Calls) Industry Standard 2017 governs how the call is conducted, and the Australian Consumer Law governs what may be represented to the person who answers. None of the three contains a rule that a caller must announce it is an AI. That absence is a finding from our own search of the instruments and of the regulator's publications as at August 2026, not a statement by a regulator that no such duty exists, and the two are not the same thing.

What does the law already require of a synthetic caller?

Everything it requires of a human one. Section 4 of the Do Not Call Register Act defines a voice call to include a call involving a recorded or synthetic voice, so a synthetic caller sits inside the definition rather than beside it. Washing, permitted hours, identification and the duty to end the call on request all follow from that single line.

What is the open question in section 9(2)?

Section 9(2) asks for the given name of the individual making the call. When the caller is software there is no individual and there is no given name, only a name somebody chose for it. The standard does not say what to do with that, and we found no ACMA guidance or determination on the point.

Two readings are available and this article is not going to pick one for you. On the first, the agency that caused the call is the relevant identity, so naming the agency and the purpose satisfies the section. On the second, giving a synthetic voice a human given name is close to what the section exists to prevent. Ask your licensee in charge which reading the agency is running on, and write the answer down.

Did the government consider an AI disclosure rule?

It considered a broader version of one and declined. In October 2025 Treasury published its review of AI and the Australian Consumer Law, concluding that the ACL is principles based and technology neutral, and it did not recommend AI specific consumer provisions. In December 2025 the government released the National AI Plan, having decided not to proceed with the mandatory guardrails it had proposed, one of which was transparency.

What that leaves is the Voluntary AI Safety Standard, which is voluntary in its name and in its effect. So the practical position for an Australian agency is that disclosure is neither compelled nor prohibited, and the decision sits with the principal rather than with a regulator.

Where has Australia written an AI disclosure rule?

In broadcasting, and only there. The Commercial Radio Code of Practice 2026 requires a licensee to provide transparency to its audience where a regularly scheduled program or news program is hosted by a synthetic voice, and it took effect on 1 July 2026. According to the ACMA that is the first time AI has been addressed in a broadcasting code of practice. It does not reach a phone call.

The contrast worth knowing is European. Article 50 of the EU Artificial Intelligence Act requires that people be informed they are interacting with an AI system unless it is obvious, and those transparency obligations began to apply on 2 August 2026. Australia has no equivalent, and Treasury's review reads as a decision not to build one.

What do Australian homeowners actually want?

To be told. Cotality surveyed buyers across four markets in early 2026 and found Australians the most sceptical of the group, which matters commercially because the people who own the houses in most farm areas are also the people most likely to want the disclosure made compulsory.

Reading those in order matters more than the headline does. 37 per cent of buyers want AI disclosure made mandatory, and that number splits hard by age: 61 per cent of Baby Boomers against 25 per cent of Gen Z. Only 40 per cent of Australian buyers would accept an AI valuation at all, the lowest of the four markets Cotality surveyed.

Distinguished Professor Jean Burgess, an Associate Director of that centre and a Chief Investigator at QUT, has described such systems as supercharging something that was already happening with endless cold calls. That is the part an agency should sit with. The reputational exposure lands on whoever the call was made for, which is the agency, not the software supplier.

How should an agency word the disclosure?

Early, in one sentence, before the pitch. Section 9(2) already forces an opening that names the business and states the purpose, so adding a few words about what is making the call costs nothing that was not already being spent.

The last line is the one that decides how the call is remembered. A person told plainly and offered a way out does not feel tricked, and it is the feeling of having been tricked, rather than the call itself, that turns into a complaint.

What should a principal actually do about it?

Disclosure is where this argument is loudest, but it is not where most agencies get caught. The obligations with penalties already attached are the ones in what the law requires of an unsolicited call, and what an AI caller can and cannot be shown to do is in does AI cold calling actually work.