The Actuaries Institute and the University of Technology Sydney’s Human Technology Institute (HTI) have released a practical guidance resource aimed at improving how financial services organisations manage risks associated with artificial intelligence.
The guidance, titled AI Risk Management in the Financial Services Sector, is positioned as a framework that can be implemented and adapted by organisations of different sizes. The organisations said the work responds to a gap between the rapid uptake of AI in financial services and the pace of governance and risk management practices.
The release cites a recent HTI survey which found that 93% of financial services organisations reported using AI, with the remainder planning to do so. However, less than half said they conducted risk assessments on internal AI use, and fewer than one in three reported AI-specific items on their risk register.
Actuaries Institute co-lead author Victor Bajanov said organisations that “do nothing in this space will be exposed to new risks they are not equipped to deal with”.
The release also points to earlier open letters from Australia’s financial services regulators, the Australian Prudential Regulation Authority (APRA) and the Australian Securities and Investments Commission (ASIC), which called on organisations to strengthen governance and risk management for AI.
The guidance highlights what it describes as “AI-amplified risks”, arguing that frontier AI systems can introduce uncertainty because they generate outputs based on patterns learned from training data, and that “agentic” AI can increase the distance between humans and system actions, reducing the opportunity for meaningful intervention.
HTI Co-Director Professor Nicholas Davis said the essential role financial services plays in society increases the need for controls to maintain confidence and stability, particularly as AI becomes embedded in decisions affecting savings, credit and insurance outcomes.
According to the release, the framework is designed to integrate with existing enterprise risk management processes rather than replace them, and is intended to remain applicable as technology evolves. It sets out guidance around four questions: governance (assigning accountability for AI risk), classification (how AI risks are categorised), quantification (how risks may be measured), and controls (appropriate controls for common AI use cases).
The release said a toolkit accompanying the framework is intended to support organisations with significant AI investments, while also being adaptable for smaller organisations or those earlier in AI deployment.
Actuaries Institute chief executive officer Elayne Grace said AI could offer benefits such as improved efficiency and service quality, but that failures could cause significant harm. She said actuaries are well placed to support decision-making as AI is embedded into operations.
You can read the full report here.
