Amazon AWS Certified Generative AI Developer - Professional AIP-C01 Practice Test 25

Question 240

A third-party analytics service wants raw model traces. What should the engineer implement at the enterprise service boundary?

Export complete prompts and sensitive tool responses because the vendor promises analytics accuracy.
Redact and minimize logged data according to sharing/purpose agreements.
Hash only user IDs while leaving confidential documents in trace attributes.
Allow the analytics service indefinite copies without expiry or deletion controls.

Explanation

Correct Answer

Redact and minimize traces before approved analytics sharing, enforcing retention, destination and contractual processing scope.

Incorrect Answers

Third-party purpose/rights restrictions still govern what data may leave the system.

The main protected content remains disclosed.

Unbounded retention violates data governance and increases risk.

Question 241

A model-neutral chat client needs consistent Bedrock message formatting. Which specific action is required by this Bedrock API behavior?

Use Converse API for supported chat models and keep a normalized message contract.
Use one model's vendor-specific InvokeModel JSON as the universal Converse request body.
Assume every Bedrock model supports the Converse API simply because its provider appears in the catalog.
Put long-lived AWS credentials into each message to simplify calling the runtime.

Explanation

Correct Answer

Use Converse only for supported message-capable models with proper role/content blocks and a normalized response envelope across providers.

Incorrect Answers

Converse has its own standardized message schema, not arbitrary provider-native fields.

Model API capabilities vary by model and version.

Credential handling belongs in IAM/authenticated backend, not chat content.

Question 242

A request uses Converse against an unsupported model/API combination. What should the integration inspect or configure at invocation time?

Force Converse for an unsupported model and assume the API translates every provider format.
Change the prompt role names only and retry without checking model capabilities.
Use a different Region at random while keeping the same unsupported API combination.
Check model API support before choosing Converse or InvokeModel.

Explanation

Correct Answer

Check the concrete model's supported inference APIs and use the appropriate Converse or model-specific InvokeModel operation.

Incorrect Answers

Unsupported model/API combinations can fail rather than silently adapt.

A schema tweak cannot enable an API that the target model does not implement.

Region changes do not necessarily change operation support.

Question 243

A chat app requires incremental text response. Which API-level change or validation correctly resolves the condition?

Invoke the nonstreaming Converse endpoint and present a fabricated token-by-token animation.
Split a final completed response into chunks after the server has already waited for full generation.
Use ConverseStream or supported streaming endpoint and handle partial deltas.
Treat every partial delta as a final validated answer before stream completion.

Explanation

Correct Answer

Use ConverseStream or another supported streaming invocation and handle typed incremental response events and final stop/error states.

Incorrect Answers

A UI animation does not deliver actual incremental model output.

Post-hoc chunking does not improve time to first real token.

A stream may terminate early with an error or missing fields.

Question 244

An app uses ConverseStream but IAM allows only InvokeModel. Which specific action is required by this Bedrock API behavior?

Grant scoped InvokeModelWithResponseStream permission for streaming.
Grant only bedrock:ListFoundationModels and expect that to authorize runtime streaming.
Use a more privileged AWS region or API key without updating least-privilege stream permissions.
Swap to a nonstreaming endpoint and claim it implements the mandatory token stream.

Explanation

Correct Answer

ConverseStream requires bedrock:InvokeModelWithResponseStream on the actual invoked resource; normal InvokeModel-only access is insufficient.

Incorrect Answers

Read-only model metadata permission cannot invoke generation.

Changing credentials blindly avoids fixing the required action/ARN scope.

It may return a final response but does not satisfy the streaming requirement.

Question 245

An ordinary Converse call receives AccessDenied. What should the integration inspect or configure at invocation time?

Verify bedrock:InvokeModel permissions and model/inference-profile ARN scope.
Grant bedrock:InvokeModelWithResponseStream only, leaving InvokeModel denied.
Add S3:GetObject wildcard permissions to fix a Bedrock invocation denial.
Disable application authentication because the service role should already be trusted.

Explanation

Correct Answer

Converse uses bedrock:InvokeModel; evaluate effective principal and the correct model or inference-profile resource permissions for AccessDenied.

Incorrect Answers

A nonstreaming Converse call requires InvokeModel permission.

Unrelated S3 permissions do not authorize model inference.

Authentication controls and Bedrock resource authorization are separate.

Question 246

A model-specific setting is absent from common inferenceConfig. Which API-level change or validation correctly resolves the condition?

Insert an arbitrary provider-specific field at the root of the Converse request.
Use supported additionalModelRequestFields only for that model.
Treat every model-specific setting as universally valid across all Bedrock models.
Turn off request validation and assume Bedrock will infer the requested vendor option.

Explanation

Correct Answer

Put supported model-specific inference settings under additionalModelRequestFields; common inferenceConfig accepts only standardized options.

Incorrect Answers

The Converse contract does not accept unknown top-level fields.

Parameter names and supported ranges vary by model/version.

Unknown fields can cause validation exceptions rather than being applied.

Question 247

A model returns a toolUse request. Which specific action is required by this Bedrock API behavior?

Have the application invoke authorized tool and supply structured toolResult.
Treat the model's toolUse proposal as proof the database update already happened.
Let a model invent the toolResult payload without contacting the backend.
Execute the tool with a shared admin role because the toolUse name is in a trusted model message.

Explanation

Correct Answer

For client-side Converse tool use, application code executes the authorized operation and returns a structured toolResult correlated to the toolUse ID.

Incorrect Answers

Tool intent precedes actual execution and may never commit.

A fabricated result cannot validate business state.

Actual actor authorization must be checked by the application.

Question 248

The Bedrock Converse model issued a toolUse request, but a subsequent toolResult does not match its tool-use identifier. Select TWO changes to the tool-response correlation and schema that repair that specific protocol error.

Generate a new toolUseId when returning the result to avoid reusing identifiers.
Preserve the tool-use request correlation ID.
Return a plain prose string in place of the required structured toolResult content block.
Return a schema-valid toolResult with the matching identifier.
Send the tool result under an unrelated prior user turn, leaving the active tool request unmatched.

Explanation

Correct Answer

The tool result must echo the original toolUseId so Bedrock can match it to the outstanding request.

A valid toolResult content block and proper message sequence are required for the multi-turn continuation.

Incorrect Answers

A mismatched ID cannot be associated with the request the model issued.

The API expects typed content, not an arbitrary narrative substitute.

Turn-level correlation and conversation structure would be incorrect.

Question 249

A tool call parameter has extra unexpected fields. Which API-level change or validation correctly resolves the condition?

Accept all additional JSON keys if the payload parses successfully.
Move extra fields into a nested metadata object and forward them to the backend unchecked.
Allow the FM to choose which schema fields to ignore at runtime.
Enforce schema validation and reject unknown/high-risk arguments.

Explanation

Correct Answer

Reject any unsupported or high-risk extra tool argument with schema/allowlist validation before executing an action.

Incorrect Answers

Parseability does not establish that unknown fields are safe or allowed.

Nesting does not eliminate unauthorized instructions or parameters.

Application/tool contract validation cannot be delegated to the model.

Question 250

A client claims a successful transfer after model generated tool intent. Which specific action is required by this Bedrock API behavior?

Display a success toast immediately after the model issues a transfer tool request.
Treat a queued asynchronous transfer as settled, with no pending state.
Wait for authoritative transaction response before asserting completion.
Generate a plausible receipt number from customer details instead of querying the ledger.

Explanation

Correct Answer

A transfer is complete only after the financial backend confirms commit and returns an authoritative transaction reference.

Incorrect Answers

Tool intent does not prove funds were moved.

Queue acceptance differs from final bank transaction status.

A fabricated identifier is not a verifiable commit receipt.

Pass Your Amazon AWS Certified Generative AI Developer - Professional AIP-C01 Exam Easy!

Question 240

Explanation

Correct Answer

Incorrect Answers

Question 241

Explanation

Correct Answer

Incorrect Answers

Question 242

Explanation

Correct Answer

Incorrect Answers

Question 243

Explanation

Correct Answer

Incorrect Answers

Question 244

Explanation

Correct Answer

Incorrect Answers

Question 245

Explanation

Correct Answer

Incorrect Answers

Question 246

Explanation

Correct Answer

Incorrect Answers

Question 247

Explanation

Correct Answer

Incorrect Answers

Question 248

Explanation

Correct Answer

Incorrect Answers

Question 249

Explanation

Correct Answer

Incorrect Answers

Question 250

Explanation

Correct Answer

Incorrect Answers

AWS Certified Generative AI Developer - Professional AIP-C01 Practice Tests