
AAIA exam practice question: daily practice for the ISACA Advanced in AI Audit (AAIA) exam — domain: AI Operations.
Question
You’re auditing an organization’s model-serving environment and find that inference endpoints accept long-lived static API keys stored in application config files. Which control most effectively reduces the risk of unauthorized model access while supporting operational scalability?
Show the answer and explanation
Correct answer: A. Issue short-lived signed tokens from an identity service with automatic rotation and revocation support for inference requests
Issuing short-lived signed tokens with automatic rotation and revocation is the most effective control: it minimizes credential exposure windows, enables centralized identity and access management, and allows immediate revocation if a credential is compromised. Mutual TLS is strong for transport authentication but annual certificate rotation is too infrequent and mutual TLS alone doesn’t provide granular, short-lived credential lifecycle or easy revocation of individual clients. Encrypting static keys and rotating them monthly still leaves long exposure and relies on manual processes that are error-prone. IP allowlisting and alerts add defense-in-depth but can be bypassed or block legitimate clients and do not reduce credential compromise risk as effectively as short-lived tokens.
Want more practice?
Prepare for the ISACA Advanced in AI Audit (AAIA) exam with AI Audit & Compliance Framework: Practical Methods & Evaluation.
