Explainable guidance
Recommendations should show the relevant assumptions, data signals and trade-offs in understandable language.
Our approach to responsible AI.
AI should improve financial understanding and decision quality without pretending to replace human judgement or regulated professional advice.
Human control
Users retain decision authority, can challenge recommendations and should be directed to qualified professionals when appropriate.
Privacy and minimisation
AI features should use the minimum necessary information and follow documented access, retention and security controls.
Fairness and testing
Models and workflows should be evaluated for bias, accuracy, harmful outcomes and performance across different user circumstances.