2.2 Determine Recommended Standard Architectures for AI Security NETWORK-SECURITY-ARCHITECT Practice Questions (Page 1)

Questions 1–5

1. 1application · medium

   A security architect is designing protection for a large language model (LLM) that is accessed via an API. The architect wants to detect prompt injection attempts and also identify when the model is being used to generate malicious code. Which combination of AI security content types should be used?

   • AAI-specific signatures and behavioral models
   • BThreat intelligence feeds and behavioral models
   • CBehavioral models and data loss prevention policies
   • DTraditional antivirus signatures and threat intelligence
   Reveal Solution Discussion Select an answer first
2. 2application · medium

   A government agency is deploying an AI system for document classification. The system is hosted in a private cloud that is not connected to the internet. The agency needs to monitor AI traffic for anomalies but does not require real-time blocking. Which AIRS form factor is most appropriate?

   • ACloud-native, using a managed service
   • BInline, deployed as a physical appliance
   • COut-of-band, using network taps to mirror traffic
   • DInline, deployed as a virtual appliance
   Reveal Solution Discussion Select an answer first
4. 3application · medium

   A startup is deploying an AI application in a Kubernetes cluster on a public cloud. The security team wants to integrate security directly into the application's deployment to inspect traffic between microservices. Which AIRS form factor should be used?

   • AInline, deployed as a physical appliance on-premises
   • BOut-of-band, using Kubernetes audit logs
   • CInline, deployed as a virtual appliance in the same VPC
   • DCloud-native, deployed as a sidecar container
   Reveal Solution Discussion Select an answer first
5. 4foundation · easy

   Why is it important to protect training data in an AI environment?

   • ATraining data often contains sensitive information that, if compromised, could lead to data breaches or model poisoning.
   • BTraining data is public and does not require protection.
   • CTraining data is only needed during development and can be discarded after deployment.
   • DTraining data is encrypted by default and cannot be accessed by attackers.
   Reveal Solution Discussion Select an answer first
6. 5application · medium

   A healthcare organization is deploying an AI application in a public cloud to analyze medical images. The security team requires that all AI traffic be inspected for malicious content and that the AI application remain highly available. The organization has no on-premises infrastructure and wants to minimize operational overhead. Which AIRS form factor is most appropriate?

   • AOut-of-band, using cloud-native logging and monitoring services
   • BInline, deployed as a virtual appliance in the same cloud VPC
   • CCloud-native, integrated as a managed service with auto-scaling
   • DInline, deployed as a physical appliance in a nearby co-location facility
   Reveal Solution Discussion Select an answer first