By Philippa Billings, Chief Advice Officer, Otivo
There isn't a special set of financial advice rules for AI in Australia. There doesn't need to be. If software crosses the line from explaining money to providing regulated financial advice, the existing financial services laws come with it. The technology may be new. The accountability isn't.
Ask an AI chatbot a question about your super and it can produce an answer in seconds. Ask whether you should invest, pay down your mortgage or put more into super and the response might be detailed, confident and remarkably personal.
But who's responsible if that answer is wrong?
That's the question that cuts through much of the noise surrounding AI and financial advice.
Australia doesn't currently have a standalone law specifically governing "AI financial advice". Instead, financial services regulation has largely been designed around what a service does, rather than the technology it uses to do it.
If regulated personal financial advice is being provided, using an algorithm doesn't make those obligations disappear.
In fact, ASIC has had specific guidance dealing with automated financial advice since 2016.
So, is AI financial advice regulated in Australia?
Yes.
Where AI is used to provide regulated financial product advice, Australia's existing financial services laws apply. ASIC describes automated financial product advice as "digital advice", also known as robo-advice or automated advice, and defines it as advice provided using algorithms and technology without the direct involvement of a human adviser.
ASIC's Regulatory Guide 255, first issued in August 2016, explains how the existing regulatory framework applies to businesses providing digital advice to retail clients.
That distinction is important.
The law isn't particularly interested in whether an answer appeared in a chat window, came from an algorithm or was delivered across a desk.
It's interested in what service was actually provided.
And that means two AI-generated answers that look almost identical on screen can sit in very different regulatory worlds.
The rules follow the advice, not the technology
That's the central principle for understanding AI financial advice in Australia.
Using new technology doesn't give a financial services business a new regulatory category.
The obligations that apply depend on the financial service being provided and the circumstances in which it's provided.
That principle has become increasingly important as AI has spread through Australia's financial system.
In October 2024, ASIC published Report 798: Beware the gap, examining the use of AI by 23 Australian financial services and credit licensees across retail banking, credit, insurance and financial advice.
ASIC analysed 624 AI use cases that were either operating or being developed as at December 2023.
Its concern wasn't that financial institutions had suddenly discovered an unregulated technology.
It was that their governance might not keep pace with how quickly they were adopting it.
ASIC found that almost half of the licensees reviewed didn't have policies addressing consumer fairness or bias in AI, and even fewer had policies dealing with disclosure of AI use to consumers. Around 60% of the licensees intended to increase their use of AI.
The technology was accelerating.
The responsibility wasn't going anywhere.
ASIC has been thinking about automated advice for a decade
It can sometimes sound as though AI has arrived faster than regulators could possibly respond.
In financial advice, that's not quite the case.
Long before ChatGPT put conversational AI in millions of people's hands, financial services businesses were already using algorithms to automate parts of the advice process.
ASIC responded with Regulatory Guide 255 in 2016.
The terminology was different. The industry generally talked about "robo-advice" or "digital advice" rather than AI financial advice.
But the fundamental regulatory problem was remarkably similar: what happens when technology, rather than a human sitting opposite the client, produces financial advice?
ASIC's answer wasn't to create an entirely separate regulatory universe for machines.
Instead, it applied the financial services framework to digital advice and added guidance addressing the particular risks created by automation.
That includes licensing, organisational competence, monitoring and testing algorithms, the quality of advice and the obligations applying when personal advice is provided.
An advice algorithm isn't supposed to be a black box
This is one of the biggest differences between simply using AI to discuss money and operating a regulated digital advice service.
ASIC expects digital advice licensees to understand, document, monitor and test the algorithms underpinning their advice.
RG 255 says providers should maintain documentation explaining the purpose, scope and design of their algorithms and have documented testing strategies.
Algorithms should be robustly tested before advice is provided and regularly afterwards.
Providers are also expected to review algorithms when circumstances change, including changes to markets or the law, keep records of changes and maintain controls that allow advice to be suspended where an algorithmic error could cause client loss or breach the Corporations Act.
ASIC also expects digital advice businesses to have adequate human and technological resources overseeing the performance of those algorithms.
Automation, in other words, doesn't mean nobody's watching.
Quite the opposite.
The most important feature may be knowing when not to answer
There's a natural temptation to judge an AI system by the number of questions it can answer.
Financial advice sometimes requires the opposite test.
Can it recognise the question it shouldn't answer?
Personal financial advice can be limited to a particular issue. Someone might want advice about their super investments without wanting a comprehensive financial plan covering every aspect of their life.
But limited advice isn't a lower regulatory standard.
ASIC says the same best-interests and related obligations apply to personal advice whether it's limited or comprehensive in scope.
That means the advice process has to recognise when the information available, or the agreed scope of the service, isn't sufficient to provide appropriate advice.
A properly designed digital advice service therefore needs boundaries.
Sometimes the correct output isn't another recommendation.
It's "I can't appropriately advise you on that here."
That isn't a failure of the technology.
It can be evidence that the system is doing its job.
What obligations apply to licensed personal financial advice?
When personal financial advice is provided, important obligations apply regardless of how impressive, simple or automated the technology behind it might be.
Section 961B of the Corporations Act 2001 contains the best interests duty. Section 961G requires resulting advice to be appropriate to the client. Section 961H deals with advice based on incomplete or inaccurate information, while section 961J requires the client's interests to be prioritised where relevant conflicts exist.
There are also disclosure and record-keeping requirements that can apply to personal advice provided to retail clients, including requirements relating to Statements of Advice and records of advice.
Then there's what happens after the advice is delivered.
Licensed financial services businesses have complaints obligations and, where required, access to external dispute resolution through the Australian Financial Complaints Authority. AFS licensees that provide financial services to retail clients are also subject to compensation-arrangement requirements under the Corporations Act.
None of those protections become less important because the customer experienced the service through a beautifully designed chat interface.
Arguably, they become easier to overlook.
Who's accountable when an algorithm gets something wrong?
Ultimately, the responsibility doesn't belong to the algorithm.
It sits with the financial services business responsible for providing the regulated service.
That's why ASIC devotes so much of its digital advice guidance to governance, testing, monitoring, record keeping and the ability to suspend faulty algorithms.
Imagine a 52-year-old receives a recommendation from a licensed digital advice service and later believes the recommendation was inappropriate because it relied on a flawed assumption.
There's an organisation standing behind that advice.
There's a complaints process. Depending on the circumstances, there may be access to AFCA. There are regulatory obligations applying to the provider.
Now imagine the same person asks a publicly available general-purpose AI tool the same question.
The answer might look just as polished.
It might even sound more convincing.
But the framework sitting behind it is completely different.
That's the distinction consumers increasingly need to understand.
What about ChatGPT and other general-purpose AI?
This is where the story becomes more complicated.
General-purpose AI tools can be enormously useful for learning about money.
They can explain compound interest. Describe how super works. Help someone understand an unfamiliar financial term. Summarise information and suggest questions someone might want to investigate.
But that doesn't make a general-purpose AI platform a licensed personal financial advice service.
ASIC's Moneysmart now specifically addresses this distinction in its guidance on AI and money decisions.
It encourages people to think of publicly available AI as a learning tool rather than a decision-maker. Moneysmart warns that AI can produce inaccurate or biased information, may not understand someone's full financial circumstances and generally isn't licensed to provide personal financial advice.
That's an important distinction because generative AI has made the experience of receiving financial information much more personal.
A chatbot remembers what you just told it. It speaks directly to you. It can incorporate your age, salary or mortgage into its answer.
The response can therefore feel like personal financial advice long before the regulatory threshold has necessarily been crossed.
Conversational fluency and regulatory accountability aren't the same thing.
The real AI advice gap isn't where you might expect
This is where Australia's technology-neutral approach is both powerful and interesting.
When a licensed financial services business uses AI to provide regulated advice, there's already a substantial regulatory framework sitting behind it.
The harder consumer problem is understanding everything outside that framework.
AI has made general financial information extraordinarily easy to generate and extraordinarily convincing to read.
The average person shouldn't need to understand the Corporations Act to work out whether the financial recommendation appearing on their phone has a regulated business standing behind it.
Yet visually, the distinction can be almost impossible to see.
One chat window can provide general information.
Another can provide licensed digital advice.
The interfaces may look almost identical.
The obligations behind them aren't.
Frequently asked questions
Is there a specific AI law for financial advice in Australia?
Not currently.
Australia regulates financial advice primarily through its existing financial services framework rather than through a standalone law specifically governing "AI financial advice".
ASIC's RG 255 then provides specific regulatory guidance for digital financial product advice.
That approach reflects an important principle: changing the technology used to deliver a regulated financial service doesn't necessarily change the underlying obligations.
Does an AI advice service have to give me a Statement of Advice?
It depends on the service and the circumstances.
The Corporations Act contains requirements relating to Statements of Advice when personal advice is provided to retail clients, along with exceptions and circumstances where other records may apply.
The important point is that those requirements are determined by the advice being provided and the applicable legal framework, not simply by whether a human or algorithm generated it.
A tool providing factual information is in a very different regulatory position from a licensed service providing personal financial advice.
Can I complain about automated financial advice?
If the advice was provided by a financial firm that's an AFCA member and the complaint falls within AFCA's jurisdiction, automated delivery doesn't remove that complaints pathway.
Your complaint is with the financial firm, not its algorithm.
That's one of the clearest practical differences between regulated digital advice and an answer generated by a general-purpose AI tool.
How can I tell whether AI financial advice is regulated?
Look beyond the chatbot.
Find out which legal entity is providing the financial service and whether the relevant service is being provided under an Australian Financial Services Licence.
Understand whether you're receiving factual information, general advice or personal advice.
Read the provider's disclosure material.
Look for its complaints process.
And ask perhaps the simplest question of all:
Who's legally standing behind this answer?
If that isn't clear, don't assume that an intelligent-sounding response carries the protections of regulated financial advice.
Where does Otivo sit?
Otivo is a licensed digital financial advice platform operating under Australian Financial Services Licence and Australian Credit Licence No. 485665.
That means the technology isn't being used as a substitute for a regulated advice framework. It operates within one.
Otivo can use a person's financial information and goals to provide advice within the areas its service is designed to cover. The platform is built around the obligations that apply to licensed advice, including recognising circumstances where the available information or scope isn't sufficient to support an appropriate recommendation.
For example, Otivo's retirement planning module can model someone's retirement position using their financial circumstances, while the super investment options module can provide advice about how their super is invested.
For circumstances that require judgement outside the scope of digital advice, a licensed human financial adviser remains the appropriate destination.
AI will almost certainly change the way Australians access financial advice.
But it doesn't change the most important question a consumer can ask.
If I act on this advice, who stands behind it?
Sources
- Australian Securities and Investments Commission, Regulatory Guide 255: Providing digital financial product advice to retail clients, issued 30 August 2016.
- Australian Securities and Investments Commission, Report 798: Beware the gap — Governance arrangements in the face of AI innovation, released 29 October 2024.
- Australian Securities and Investments Commission, ASIC warns governance gap could emerge in first report on AI adoption by licensees, 29 October 2024.
- Australian Securities and Investments Commission, Moneysmart, AI and money decisions, last updated 17 August 2026.
- Australian Securities and Investments Commission, Tips for giving limited advice.
- Corporations Act 2001 (Cth), ss 946A, 961B, 961G, 961H and 961J.
- Australian Securities and Investments Commission, Market Integrity Update, Issue 177, June 2026.
Disclaimer
The information in this communication is current as at September 2026 and has been prepared by Otivo Pty Ltd ABN 47 602 457 732, AFSL and Australian Credit Licence No. 485665. This content is general information only and has been prepared without taking into account your objectives, financial situation or needs. It is not personal financial or taxation advice and should not be relied on as such. Before acting on any information, you should consider its appropriateness having regard to your personal circumstances. This material must not be reproduced in whole or in part, or posted on any social media platform, without the prior written consent of Otivo Pty Ltd.