The member best interests duty and AI
The Superannuation Industry (Supervision) Act's member best interests duty — which requires trustees to act in members' best financial interests — is the foundational governance principle for superannuation AI. Every AI system deployed by a superannuation fund must be assessed against this duty: does this AI system, as deployed, serve the best financial interests of members? This question has specific content for each AI use case.
For investment AI — models used in asset allocation, risk management, manager selection, or factor exposure — the best interests duty requires that the AI produces better investment outcomes for members than available alternatives, that it is properly governed and monitored, and that investment staff understand its outputs and limitations sufficiently to exercise genuine investment judgment. An AI model that investment staff treat as a black box, accepting its outputs without critical evaluation, raises questions about whether the trustee is genuinely exercising its investment governance obligations.
AI in member communications: the advice boundary
Many superannuation funds are deploying AI in member communications — chatbots that answer member enquiries, personalised retirement projections, AI-generated consolidation recommendations, and personalised investment option suggestions. The regulatory challenge is that these communications can cross the boundary from factual information into financial advice, which requires an Australian Financial Services Licence and compliance with the best interests duty and appropriate advice obligations.
The ASIC guidance on scaled and digital advice is directly relevant. AI systems that take into account a member's specific circumstances and make personalised recommendations about financial products or strategies are providing financial advice regardless of whether a human advisor is involved. Funds must assess their AI communication tools against this boundary and ensure that tools that cross into advice territory are governed as advice tools — with appropriate authorisations, compliance frameworks, and member consent arrangements.
Claims processing AI and member rights
AI used in insurance claims processing — assessing death benefit claims, total and permanent disability claims, and income protection claims — affects members at the most difficult moments of their lives. The governance obligations for claims AI are particularly demanding because the stakes are high and because the members affected are often in vulnerable circumstances. ASIC has specifically addressed the obligation of superannuation trustees to handle claims efficiently, honestly, and fairly — an obligation that applies to AI-assisted claims processes as much as to human processes. The key requirements: claims decisions influenced by AI must be explainable to the member in plain language, there must be a genuine human review pathway, and the AI system must be monitored for systematic errors that disadvantage members.
