AI in Regulatory Compliance Quiz

Answers and explanations appear as you take the quiz.

1. What can machine-learning triage help reduce in transaction monitoring?
   • Lower-priority or false-positive alerts that require manual review
   • The need to verify a customer’s identity
   • The number of laws that apply to a bank
   • The need to investigate any concerning activity
2. What does KYC stand for in a compliance context?
   • Keep Your Cash
   • Know Your Customer
   • Key Yield Calculation
   • Know Your Controls
3. How can transaction-graph analysis help an AML team?
   • It proves every connected account is criminal
   • It replaces customer due diligence
   • It can reveal connected accounts and fund-flow patterns that isolated transactions may not show
   • It automatically files a final legal conclusion
4. What is a useful role for retrieval when an LLM answers a regulatory question?
   • Guarantee the model has interpreted the law correctly
   • Replace checking the effective date and jurisdiction
   • Automatically determine legal compliance
   • Surface relevant source passages for a reviewer, while leaving the answer open to verification
5. What does SupTech refer to?
   • Technology used by financial authorities to support supervision
   • A consumer budgeting app
   • An automated court judgment system
   • A type of bank account
6. Why classify an AI system’s specific use before applying EU AI Act duties?
   • All AI systems receive identical requirements
   • Requirements depend on its intended use, risk category, and applicable dates
   • The model’s architecture alone determines every obligation
   • A compliance label removes the need for legal review
7. Why should a compliance retrieval system track the source document’s version and effective date?
   • Versioning makes model outputs legally binding
   • An effective date proves a model is unbiased
   • A retrieved passage may be outdated or not yet applicable to the use case
   • Old rules always remain applicable everywhere
8. What is an appropriate role for AI transaction-monitoring alerts?
   • Declare a customer guilty without investigation
   • Replace all suspicious-activity reporting controls
   • Automatically close every low-score alert without governance
   • Help prioritize cases for review while analysts assess evidence and context