AI Regulation Tracker / Regulator guidance
ASIC: AI does not change a financial adviser's best-interests duty or complaint-handling obligations
Australia's corporate regulator is telling AFS licensees that using AI to help produce advice changes none of their existing legal duties. The adviser and the licensee still own the recommendation in full.
The Australian Securities and Investments Commission has again made clear that using artificial intelligence in financial advice does not shift, soften, or suspend an adviser's legal duties. Its Financial advice update of February 26, 2026 sets out the obligations ASIC expects licensees to meet, including the best interests duty and internal dispute resolution requirements, and it does so without creating any separate track for automated or AI-assisted advice. The message for firms rushing to deploy advice tools is direct. The obligations that applied to a human adviser apply to the same adviser using a model.
The regulator's position is tech-neutral
ASIC has not written a bespoke rulebook for AI in advice. Instead it applies the existing framework. As ASIC Chair Joe Longo put it when the regulator published its first review of AI adoption, "existing consumer protection provisions, director duties and licensee obligations put the onus on institutions to ensure they have appropriate governance frameworks." That framing matters. It means an AFS licensee cannot point to a vendor's algorithm, or to the absence of an AI-specific statute, to explain a recommendation that fails the client. The licensee and the adviser own the output.
What the February update actually covers
The February 2026 update itself is a compliance briefing rather than an AI announcement. It addresses internal dispute resolution reporting, self-managed super fund establishment advice, reportable situations, offshore outsourcing, and recent enforcement and panel decisions. Its relevance to AI is structural. ASIC's complaint-handling standard, set out in Regulatory Guide 271, treats any expression of dissatisfaction as a complaint that must be captured and resolved, and that duty does not change when part of the advice was produced by a tool. A firm that lets an AI system field client questions still has to log, handle, and report complaints the same way.
Where ASIC has spelled out the AI risk
The clearest statement of ASIC's thinking sits in Report 798, "Beware the gap: Governance arrangements in the face of AI innovation," published on October 29, 2024. Reviewing 23 licensees, ASIC found AI use was still relatively cautious but accelerating, with about 60 percent of the group planning to increase it. The regulator warned that nearly half lacked policies addressing consumer fairness or bias. Longo cautioned that "without appropriate governance, we risk seeing misinformation, unintended discrimination or bias, manipulation of consumer sentiment and data security and privacy failures." The report also noted that even fewer licensees had policies on disclosing AI use to consumers, which leaves clients unaware that a tool shaped their advice. APRA has since called for a step-change in how regulated entities manage AI risk, describing governance and controls that are not keeping pace with the speed and scale of adoption. Advisers therefore face a consistent signal from both the prudential and the conduct regulator, rather than mixed messages they could play against each other.
The duties that do not bend for automation
Three obligations do the work here. The best interests duty requires an adviser to act in the client's best interests and to base a recommendation on the client's objectives, situation, and needs. The obligation to provide financial services efficiently, honestly and fairly sits on the licensee at all times. And the internal dispute resolution rules govern how client dissatisfaction is handled. An adviser who cannot explain how an AI tool reached a conclusion has a practical problem, because it is difficult to show a recommendation meets the best interests duty when no one can account for how it was formed. The same logic applies to the complaint-handling rules. If a client disputes AI-assisted advice, the firm must be able to reconstruct and justify the recommendation through its internal dispute resolution process, and later to the Australian Financial Complaints Authority if the matter escalates. Explainability is therefore not a technical nicety. It is how a firm demonstrates compliance and defends its file.
What it does not do
This posture does not ban AI in advice, and it does not require firms to disclose every tool in every document. It does not create a new licence condition or a new penalty. What it removes is the argument that automation dilutes responsibility. For a United States reader, the parallel is close. Australia's tech-neutral approach mirrors the way US regulators apply fiduciary duty and existing conduct rules to AI, and it previews the standard advisers operating across both markets should expect: the human and the firm answer for the machine's output.
Frequently Asked Questions
What did ASIC actually change for AI in financial advice?
Nothing was carved out. ASIC's February 2026 advice update restates existing duties, including the best interests duty and internal dispute resolution rules, and applies them on a tech-neutral basis. Read with Report 798 on AI adoption, the position is that current obligations govern AI use. No AI-specific advice rule was created.
Who is affected by this?
AFS licensees and their authorised financial advisers that use or plan to use AI in advice, research, file notes, complaint handling, or client triage, and the advice-technology vendors that supply those tools to licensees.
Does using AI reduce an adviser's responsibility for a recommendation?
No. ASIC's stated position is that existing consumer protection, director, and licensee obligations put the onus on the firm. The licensee and the adviser own an AI-assisted recommendation in full, the same as one produced without a tool.
Do advisers have to be able to explain how the AI reached its conclusion?
In practice, yes. A firm must be able to show a recommendation meets the best interests duty. If no one can account for how an AI output was formed, demonstrating that the advice serves the client's interests becomes difficult, which is why explainability matters for compliance.
Does this bind US firms?
Not directly. It governs Australian AFS licensees and advisers. For US firms it is a useful preview, because it aligns with how US regulators apply fiduciary duty and existing conduct rules to AI rather than writing separate AI advice rules.
Sponsored Training
Browse the full AI Regulation News tracker
Informational analysis for working professionals, not legal advice. Confirm how any rule applies to your situation with qualified counsel.