AI Safety, Security, and Governance Questions
Practice questions for AI Safety, Security, and Governance topic in AWS Certified Generative AI Developer - Professional. 40 questions covering this domain.
A company uses a self hosted model outside Amazon Bedrock but wants to apply the same Guardrails policies to its inputs and outputs. Which approach is...
A security architect needs private network connectivity from an Amazon VPC to Amazon Bedrock without sending traffic over the public internet. Which A...
A governance team wants real time control over which tools an agent may call, using natural language policies that can be converted to Cedar. Which Ag...
A procurement lead asks which statement about Amazon Bedrock customer content privacy is accurate. Which answer should the architect provide?
A legal assistant uses retrieved case materials, and the firm wants a safeguard that checks whether generated answers stay grounded in the provided so...
A central platform team wants all Bedrock inference calls in member accounts to enforce a standard guardrail automatically without each application sp...
An enterprise agent must securely access AWS services and third party tools on behalf of users after pre authorization. Which AgentCore capability is ...
A compliance team wants to proactively block questions and responses about topics the company has declared off limits. Which Guardrails control is des...
A bank wants to automatically detect and mask Social Security numbers and custom employee identifier patterns in prompts and responses. Which Guardrai...
A team is concerned about jailbreaks, prompt injection, and prompt leakage attempts against its generative AI assistant. Which Guardrails capability d...
Which Bedrock Guardrails policy targets prompt-injection and jailbreak attempts specifically?
What is the principal purpose of Bedrock Guardrails contextual grounding checks?
An organization runs a non-Bedrock self-hosted LLM but wants to apply the same policies (denied topics, PII redaction) to its inputs and outputs as a ...
Which AWS service catalogs sensitive data discovered in Amazon S3 (for example, in a RAG document store) so governance teams can tag and protect it be...
A regulator asks an enterprise to demonstrate that prompts and responses are not used by AWS to train base foundation models. Which AWS-published stat...
Which Amazon Bedrock Guardrails capability detects and redacts or blocks specific PII like email addresses or phone numbers in prompts and responses?
A platform team must ensure all Bedrock InvokeModel calls in member accounts include a guardrailIdentifier. Which combination most reliably enforces t...
Which Amazon Bedrock Guardrails control policy is designed to detect and block harmful content categories such as hate, insults, sexual content, or vi...
An organization wants centralized AI risk management policies applied across many accounts including model approvals, ML lineage, and risk register in...
Which AWS feature should be used to record and audit every Bedrock control plane API call (such as CreateAgent or PutFoundationModelEntitlement) made ...
Sign in to see all 40 questions
Create a free account to browse all questions — completely free during our launch phase.