AI Compliance & Regulation
ASIC's 2027 Priorities: Your AI Must Let Customers Reach a Human
ASIC named AI and automated-system harm one of nine 2027 enforcement priorities. What brokers, advisers and insurance brokers using AI should fix first.
In short. Yes, ASIC will act on AI failures in 2027. On 18 November 2026 it named consumer harm from poorly implemented AI and automated systems as one of nine enforcement priorities. The focus is systems that give misleading information, make inappropriate decisions or block customers from help. A licensee using AI should make sure a customer can reach a human for complaints, scams, claims and hardship requests.
Does software help decide things about your customers or staff? From 10 December 2026 your privacy policy may have to say so. Six questions give a likely answer.
Take the check →ASIC just put AI on its enforcement list
On 18 November, ASIC named consumer harm from "poorly implemented technologies, including AI and automated systems" as one of its nine enforcement priorities for 2027. If you hold an AFS or credit licence and a chatbot, an automated triage tool or an AI-drafted letter sits between you and your clients, you are now inside a named target area.
ASIC Chair Sarah Court put it plainly: "The technology may be new. The responsibility is not."
What ASIC actually announced
The 2027 enforcement priorities were released at ASIC's Annual Forum. Cost-of-living misconduct by banks, insurers and super trustees leads the list. Harmful financial advice and lead generation models are on it. So is the AI item, which ASIC frames as "serious failures in the way financial firms use AI and other technology".
ASIC names three failure types: systems that produce misleading information, systems that make inappropriate decisions, and automated systems that put barriers between customers and the help they need. Court singled out one safeguard. Human review, she said, is "particularly important" when customers make complaints, respond to scams, pursue insurance claims or seek financial assistance.
None of this is a new law. Court was explicit that AI, automated decision-making and chatbots do not change a firm's legal obligations. What has changed is where ASIC is pointing its investigators.
1 of 9
ASIC 2027 enforcement priorities targeting AI and automated systems
Announced 18 Nov 2026
4
Moments where ASIC says human review matters most
Complaints, scams, claims, hardship
Why small brokers and advisers should read this as their problem
The headline names banks and insurers, so it's easy to file this under "big end of town". That would be a mistake. The AI priority is written about financial firms generally, and the tools small licensees are adopting now — website chatbots, automated lead qualification, AI-drafted file notes and claims correspondence — are exactly the systems ASIC is describing.
ASIC has been signalling this for two years. Its 2024 review of 23 licensees, Report 798, warned that AI use was at risk of outrunning the governance around it. Fourteen of those licensees planned to increase their AI use. We covered the regulator's shift to AI accountability in its corporate plan in September. The enforcement priority is the next step: from warning to investigation.
The real risk for a small practice isn't the AI giving a bad answer once. It's a chatbot that quietly loops a distressed client away from a person for weeks — a hardship request that never reaches you, a claim query that dead-ends. That is a revenue capture problem as much as a compliance one: the client you fail at the worst moment is the client who leaves. If you'd like a hand checking where your own tools sit, book a short call.
ASIC's own test for when it will litigate
Court also said what won't trigger litigation. Naming a priority, she said, "does not create a presumption that every failure will lead to litigation". ASIC will distinguish isolated failures that are promptly fixed from misconduct that is extended, repeated or systemic, and it will weigh remediation, cooperation and whether known risks raised with senior leaders were ignored.
Read that as a practical instruction. A firm that has mapped its AI touchpoints, built a human escalation path and fixes problems fast is in a very different position from one that switched on a chatbot and never looked at the transcripts.
What to do before 2027 starts
Four steps, most of which take an afternoon. List every place AI or automation touches a client: chatbot, enquiry triage, document drafting, claims or complaints routing. For each one, test the four moments ASIC named — complaint, scam, claim, hardship — and confirm a client can reach a person quickly and obviously. Keep a record of who reviews the AI's output and how often. And check your privacy policy: from 10 December 2026, the Privacy Act's automated decision disclosure rule also applies to covered businesses.
None of this is a reason to pull back from AI. Brokers and advisers using it well are clearing admin that used to eat their evenings. The firms that come out ahead in 2027 will be the ones that can show the human is still in the loop where it counts. Our AI services for mortgage brokers, financial advisers and insurance brokers start with exactly that map.
Key takeaways
Common questions
Is ASIC targeting AI in 2027?
Yes. ASIC's 2027 enforcement priorities, announced on 18 November 2026, include consumer harm from poorly implemented technologies, including AI and automated systems. The focus is systems that give misleading information, make inappropriate decisions or stop customers getting help.
Do financial services chatbots have to offer a human in Australia?
ASIC hasn't created a new rule, but its Chair said human review is particularly important when customers make complaints, respond to scams, pursue insurance claims or seek financial assistance. Existing obligations still apply when AI is used, so a chatbot that blocks those paths is a risk.
Will ASIC sue a small firm for one AI mistake?
ASIC says a priority does not mean every failure leads to litigation. It will distinguish isolated, promptly remedied failures from extended, repeated or systemic misconduct, and will take remediation and cooperation into account.
Sources
Assumptions & methodology
- The nine priorities, the AI priority wording, the four human-review situations and the quotes from ASIC Chair Sarah Court are from ASIC media release 26-286MR (18 November 2026). The human-review situations, the litigation-threshold remarks and the closing quote are from Sarah Court's published speech at the ASIC Annual Forum (18 November 2026).
- ASIC REP 798 "Beware the gap" was published 29 October 2024 and reviewed 23 AFS and credit licensees, 14 of which planned to increase their AI use.
- The four-step checklist is CoterieLabs' practical reading of ASIC's stated focus, not ASIC guidance. Last reviewed 20 November 2026 · general information, not legal or compliance advice.
If this applies to you
Would your AI pass ASIC's human-review test?
We map where AI touches your clients, test the four moments ASIC named, and set up the escalation and review records that show the human is still in charge. Book a call to start with your own tools.