Risk & governance
Algorithmic fairness
Assessing whether model outcomes treat relevant groups appropriately.
Definition
Algorithmic fairness uses legal, ethical and statistical criteria to evaluate differences in errors, decisions or benefits across populations.
Why it matters
No single fairness metric fits every banking decision, so legal context, customer impact and trade-offs must be explicit.
Related concepts
- AI bias
Systematic differences in model behaviour that may create unfair outcomes.
- False positive
A system flags something as present or risky when it is not.
- False negative
A system misses something that is genuinely present or risky.