Uncertainty Quantification for Financial Foundation Models: A Survey
How AI models should track and explain their uncertainty about financial decisions
Financial AI systems now handle everything from answering questions to making investment recommendations, but they rarely track where their uncertainty actually comes from. This survey identifies six distinct sources of uncertainty—from dodgy input data to flawed reasoning to unpredictable market outcomes—and shows that most current systems only measure one piece of the puzzle rather than following uncertainty through an entire financial workflow.
When an AI recommends a trade or interprets a company's financial report, you need to know not just whether it's confident, but why—and whether that confidence is justified. Right now, most financial AI systems behave like a doctor who tells you their diagnosis is 95% certain without explaining whether that certainty comes from clear test results, ambiguous symptoms, or just pattern-matching. This framework could help banks, investors, and regulators actually understand when to trust an AI's output and when to demand human review.