How AI can help people make better financial decisions: Improving access to financial help, guidance and advice
This Paper sets out four practical areas for action that could help Australia fulfil the promise of AI-enabled financial advice and combat the shortcomings of existing chatbots and AI tools.
Most Australians make significant financial decisions with little or no support. Adviser numbers have fallen by almost half since 2018 to about 15,000, and only around one in six Australians over 55 receives financial advice. This Discussion Paper asks whether AI can help close that gap.
In developing this Paper, the Actuaries Institute's AI and Advice Working Group tested four popular chatbots using five hypothetical Australian retirement and investment scenarios. The authors assessed the responses and found that while the chatbots made some acceptable recommendations, most responses were potentially incomplete or potentially inappropriate.
"What is most concerning is that all four failed the most consequential question in our testing – a detailed query about retirement advice – at a time when one in eight over-45s use AI for super and retirement questions."
— Nathan Bonarius, Public Policy Council Committee member
The Paper proposes a roadmap for responsible AI-enabled help, guidance and advice:
- Raise financial capability, digital literacy and AI literacy, including updating the National Financial Capability Strategy with measurable targets and considering an AI-powered chatbot on MoneySmart as a trusted alternative to general-purpose chatbots.
- Encourage Australian innovation in financial services, including better-supported regulatory sandboxes for trialling new AI services.
- Update regulation and enforcement to clarify and enforce rules on unlicensed AI advice, update digital advice guidance and resolve uncertainty around nudges and prompts, with consistent licensing, complaints processes and accountability whether advice is delivered by people or AI.
- Strengthen regulators’ AI capabilities, including ASIC using AI to analyse publicly available data to identify potential breaches and prioritise high-risk providers.
The paper suggests that as technology and regulation evolve to overcome the barriers to AI-enabled advice, there could be greater convergence between factual information, general advice and personal advice.
"Progress will depend on working together to lift consumers' financial and AI literacy, give responsible providers room to innovate, and strengthen regulation and supervision so consumers are protected. Actuaries can help support innovation and build trust by designing and testing AI-enabled advice, drawing on their expertise in modelling, risk and governance, with a focus on good outcomes for consumers."
— Elayne Grace, Actuaries Institute CEO
This Paper is authored by Tim Furlan, Kandice Cai, Ean Chan, Peter Cominos, Mudit Gupta, Tim Lam, Estelle Liu, Anthony Saliba and Edward Tam from the Actuaries Institute's AI and Advice Working Group.
This paper builds on a previous Actuaries Institute Discussion Paper, Financial Advice Reform and Help, Guidance and Advice , and follows the release of the Actuaries Institute and UTS Human Technology Institute paper AI Risk Management for the Financial Services Sector.
Related publications
Be informed. Stay ahead. Subscribe.
Receive industry-leading perspectives, straight to your inbox.
