AI Regulation Tracker  /  Regulator review

New Zealand's FMA plans a thematic review of AI in financial advice and says conduct duties still apply

New Zealand's Financial Markets Authority says AI can widen access to financial advice but must sit inside existing conduct duties, and it plans a thematic review of how Financial Advice Providers use it.

By Anthony Guerriero  ·  Reviewed by The Leveraged Years Editorial Desk  ·  Published July 7, 2026  ·  Last updated July 7, 2026

The Financial Markets Authority (FMA) has set out how it views artificial intelligence in regulated financial advice, and its message to Financial Advice Providers (FAPs) is that the technology is welcome but the rules have not moved. In its March 2026 report, "Access to financial advice in New Zealand: Challenges and opportunities for the financial advice sector," the regulator examined how firms use digital advice and AI, and said any such use must be designed with oversight that keeps existing conduct obligations "managed and met." The FMA also confirmed it is "planning to undertake a thematic review of the use of AI in financial advice."

What the FMA actually said

The report is a review, not a new rule. It draws on 80 interviews and questionnaire responses from 30 FAPs, and it treats AI as an opportunity to widen access to advice rather than a threat to be shut down. The FMA reported that, on 2025 regulatory returns, 40 FAPs run a digital advice facility, and that roughly 165,000 retail clients received digital advice in the past year, which it described as a 90 percent increase on the prior period. It found that the predominant use of AI in the sector is "to enable rather than replace advisers," with adviser oversight and accountability retained. It also noted that many FAPs are taking a cautious approach to adopting AI, which it read as room to build confidence in combining innovation with conduct obligations.

The regulator did not claim AI is trusted by consumers today. Its own consumer research put trust scores for AI-provided advice between 28 percent and 41 percent across the products tested, which the FMA said leaves confidence far from universal.

The duties that still apply

Nothing in the report loosens the conduct regime. Advisers and FAPs remain bound by the Code of Professional Conduct for Financial Advice Services. The report points specifically to Code Standard 3, which requires advice to be suitable and supports record-keeping, and Code Standard 4, which requires reasonable steps to ensure the client has understood the advice. Disclosure duties under the Financial Markets Conduct Act 2013 and its regulations continue to apply. The FMA's position is that these duties do not change because an algorithm produced part of the advice. If AI drafts a fact find, summarises a client interaction, or generates a recommendation, the FAP still has to show the advice was suitable, that the client understood it, that required disclosures were made, and that records exist.

That is where the practical risk sits. AI tools can produce fluent output that is wrong, so a firm that relies on them without checking exposes itself under duties that already bind it. The report frames the fix as design and oversight, not new law.

The line between advice and information

The review also marks the boundary the regime already draws. Providing regulated financial advice in New Zealand requires a licensed FAP under the Financial Markets Conduct Act 2013. The FMA noted that some fintech firms it engaged did not hold an FAP licence, and that while some wanted to provide advice directly, they felt constrained by the regime. Others kept their outputs "high level or generic" so they would remain information-only and not be misconstrued as financial advice. The implication for AI vendors is direct. A tool that crosses from information into a personalised recommendation is giving regulated advice, and the licensing and conduct obligations attach whether a human or a model produced it.

What the review does not do

The report does not create AI-specific rules, set an AI compliance deadline, or approve any particular tool. It signals a thematic review to come and invites the sector to share real-world use cases. For US firms, it carries no direct force. Its value is as a read on direction. New Zealand's principles-based approach, applying existing conduct duties to AI rather than writing separate ones, broadly parallels the posture Australia's ASIC has taken across the Tasman, and it previews how principles-based regulators are likely to treat AI-assisted advice, as conduct that must meet the standards already on the books.

For a FAP, the safe assumption is that the thematic review will look for exactly what the current duties require. Firms that can already show where AI sits in their process, and how oversight, suitability, disclosure, and record-keeping are handled, will have little to fear from it.

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.