Select an option, then click Submit answer.
Correct answer: Enabling stack call and debugging level traces at the function level
D When a security engineer's primary goal is to gain visibility into the internal activity of an AI-based system during runtime, they need to peer inside the "black box". AI applications and machine learning pipelines involve highly complex, multi-layered, and non-linear computations. Relying on superficial or external logs won't show you how a model arrived at a decision or where an internal pipeline failure occurred.
D Explanation The question states that the security engineer is primarily concerned about the visibility of internal activity within an AI-based system during runtime operations. Stack calls and debugging-level traces provide detailed visibility into: Function execution paths Internal model interactions Component-to-component communication Error conditions and execution flow Runtime behavior of the application This level of telemetry is essential for understanding what is happening inside the AI system, making it the best choice for monitoring internal operations.
D The security engineer's specific concern is gaining visibility into the internal activity of the AI system during runtime operations.Function-level stack calls and debugging traces log the explicit execution path, variable states, and function arguments inside the application logic. This gives the engineer granular, white-box visibility to see exactly how data flows internally and precisely where a failure, logic bypass, or prompt injection exploit is occurring during runtime. SIEM collects and correlates security events from logs — provides visibility of events, not internal function-level execution details
D The key phrase here is "visibility of internal activity." Stack call and debug-level traces give you the deepest look into what's actually happening inside the AI system at runtime — function calls, execution flow, internal state — which is exactly what you need for that level of observability.
A A SIEM aggregates logs from multiple systems, correlates events, and provides centralized visibility. Best suited for runtime monitoring + internal activity visibility.