AI Security & Compliance interview questions

1. A.mutual TLS where client and server both present certificates, authenticating service-to-service traffic
2. B.masking, encrypting or tokenizing personally identifiable information to meet privacy regulation
3. C.storing credentials in a vault and injecting them at runtime instead of hardcoding them
4. D.granting an identity only the minimum permissions needed, limiting blast radius if compromised

Answer + AI explanation with Pro

1. A.Prompt injection
2. B.Least privilege
3. C.Secrets management
4. D.Network segmentation

Answer + AI explanation with Pro

1. A.Least privilege — mutual TLS where client and server both present certificates, authenticating service-to-service traffic
2. B.Least privilege — an auditable record of which data and code produced a model, required for regulated AI governance
3. C.Least privilege — an attack where crafted input overrides an LLM's instructions to leak data or perform unintended actions
4. D.Least privilege — granting an identity only the minimum permissions needed, limiting blast radius if compromised

Answer + AI explanation with Pro

1. A.masking, encrypting or tokenizing personally identifiable information to meet privacy regulation
2. B.granting an identity only the minimum permissions needed, limiting blast radius if compromised
3. C.role-based access control granting permissions to roles that users or service accounts are bound to
4. D.storing credentials in a vault and injecting them at runtime instead of hardcoding them

Answer + AI explanation with Pro

1. A.PII handling
2. B.Secrets management
3. C.RBAC
4. D.Model and data lineage

Answer + AI explanation with Pro

1. A.Secrets management — isolating workloads into separate subnets/zones so a breach in one cannot reach the others
2. B.Secrets management — masking, encrypting or tokenizing personally identifiable information to meet privacy regulation
3. C.Secrets management — granting an identity only the minimum permissions needed, limiting blast radius if compromised
4. D.Secrets management — storing credentials in a vault and injecting them at runtime instead of hardcoding them

Answer + AI explanation with Pro

1. A.masking, encrypting or tokenizing personally identifiable information to meet privacy regulation
2. B.isolating workloads into separate subnets/zones so a breach in one cannot reach the others
3. C.mutual TLS where client and server both present certificates, authenticating service-to-service traffic
4. D.an auditable record of which data and code produced a model, required for regulated AI governance

Answer + AI explanation with Pro

1. A.Network segmentation
2. B.Secrets management
3. C.mTLS
4. D.Prompt injection

Answer + AI explanation with Pro

1. A.mTLS — an auditable record of which data and code produced a model, required for regulated AI governance
2. B.mTLS — mutual TLS where client and server both present certificates, authenticating service-to-service traffic
3. C.mTLS — an attack where crafted input overrides an LLM's instructions to leak data or perform unintended actions
4. D.mTLS — masking, encrypting or tokenizing personally identifiable information to meet privacy regulation

Answer + AI explanation with Pro

1. A.an attack where crafted input overrides an LLM's instructions to leak data or perform unintended actions
2. B.masking, encrypting or tokenizing personally identifiable information to meet privacy regulation
3. C.role-based access control granting permissions to roles that users or service accounts are bound to
4. D.an auditable record of which data and code produced a model, required for regulated AI governance

Answer + AI explanation with Pro

1. A.Least privilege
2. B.Model and data lineage
3. C.mTLS
4. D.RBAC

Answer + AI explanation with Pro

1. A.RBAC — isolating workloads into separate subnets/zones so a breach in one cannot reach the others
2. B.RBAC — role-based access control granting permissions to roles that users or service accounts are bound to
3. C.RBAC — storing credentials in a vault and injecting them at runtime instead of hardcoding them
4. D.RBAC — an auditable record of which data and code produced a model, required for regulated AI governance

Answer + AI explanation with Pro

1. A.an auditable record of which data and code produced a model, required for regulated AI governance
2. B.storing credentials in a vault and injecting them at runtime instead of hardcoding them
3. C.masking, encrypting or tokenizing personally identifiable information to meet privacy regulation
4. D.mutual TLS where client and server both present certificates, authenticating service-to-service traffic

Answer + AI explanation with Pro

1. A.Network segmentation
2. B.PII handling
3. C.Least privilege
4. D.mTLS

Answer + AI explanation with Pro

1. A.PII handling — mutual TLS where client and server both present certificates, authenticating service-to-service traffic
2. B.PII handling — masking, encrypting or tokenizing personally identifiable information to meet privacy regulation
3. C.PII handling — an auditable record of which data and code produced a model, required for regulated AI governance
4. D.PII handling — storing credentials in a vault and injecting them at runtime instead of hardcoding them

Answer + AI explanation with Pro

1. A.mutual TLS where client and server both present certificates, authenticating service-to-service traffic
2. B.an auditable record of which data and code produced a model, required for regulated AI governance
3. C.isolating workloads into separate subnets/zones so a breach in one cannot reach the others
4. D.an attack where crafted input overrides an LLM's instructions to leak data or perform unintended actions

Answer + AI explanation with Pro

1. A.mTLS
2. B.Least privilege
3. C.Prompt injection
4. D.Network segmentation

Answer + AI explanation with Pro

1. A.Network segmentation — role-based access control granting permissions to roles that users or service accounts are bound to
2. B.Network segmentation — isolating workloads into separate subnets/zones so a breach in one cannot reach the others
3. C.Network segmentation — mutual TLS where client and server both present certificates, authenticating service-to-service traffic
4. D.Network segmentation — storing credentials in a vault and injecting them at runtime instead of hardcoding them

Answer + AI explanation with Pro

1. A.an attack where crafted input overrides an LLM's instructions to leak data or perform unintended actions
2. B.storing credentials in a vault and injecting them at runtime instead of hardcoding them
3. C.mutual TLS where client and server both present certificates, authenticating service-to-service traffic
4. D.an auditable record of which data and code produced a model, required for regulated AI governance

Answer + AI explanation with Pro

1. A.Model and data lineage
2. B.RBAC
3. C.mTLS
4. D.Prompt injection

Answer + AI explanation with Pro

1. A.Prompt injection — storing credentials in a vault and injecting them at runtime instead of hardcoding them
2. B.Prompt injection — role-based access control granting permissions to roles that users or service accounts are bound to
3. C.Prompt injection — an attack where crafted input overrides an LLM's instructions to leak data or perform unintended actions
4. D.Prompt injection — an auditable record of which data and code produced a model, required for regulated AI governance

Answer + AI explanation with Pro

1. A.granting an identity only the minimum permissions needed, limiting blast radius if compromised
2. B.an auditable record of which data and code produced a model, required for regulated AI governance
3. C.role-based access control granting permissions to roles that users or service accounts are bound to
4. D.storing credentials in a vault and injecting them at runtime instead of hardcoding them

Answer + AI explanation with Pro

1. A.RBAC
2. B.Secrets management
3. C.Model and data lineage
4. D.Network segmentation

Answer + AI explanation with Pro

1. A.Model and data lineage — role-based access control granting permissions to roles that users or service accounts are bound to
2. B.Model and data lineage — masking, encrypting or tokenizing personally identifiable information to meet privacy regulation
3. C.Model and data lineage — an auditable record of which data and code produced a model, required for regulated AI governance
4. D.Model and data lineage — mutual TLS where client and server both present certificates, authenticating service-to-service traffic

Answer + AI explanation with Pro