AIF-C01 Practice Test 44 — 15 Questions

Question 1

A company is using Amazon Fraud Detector to detect fraudulent transactions. Which TWO actions can be taken to improve model accuracy? (Select TWO.)

Accepted Answer

Increase the volume of event data. Option A is correct because Amazon Fraud Detector's model accuracy improves with more historical event data — a larger volume of labeled fraud/legitimate events gives the training algorithm more examples to learn patterns from, reducing overfitting and improving predictions. Option E is correct because model accuracy depends heavily on feature quality; choosing event variables with stronger predictive power (e.g., customer age, order price, IP address, email domain) directly improves the model's ability to distinguish fraudulent from legitimate transactions. Option B is incorrect because deploying a model to multiple endpoints only affects availability/scalability of predictions, not the underlying model's accuracy. Option C is incorrect because changing the detector type (e.g., online fraud, transaction fraud) alters the use case rather than inherently improving accuracy for the existing scenario. Option D is incorrect because switching to a different model version simply selects an already-trained model; it does not by itself improve accuracy unless retraining with better data or variables occurs.

Answer

Deploy the model to multiple endpoints

Answer

Use a different detector type

Answer

Use a different model version

Question 2

Which THREE are benefits of using Amazon Bedrock over self-managing foundation models on EC2? (Choose THREE.)

Accepted Answer

Built-in integration with AWS services such as AWS CloudWatch and AWS CloudTrail.. Option A is correct because Amazon Bedrock natively integrates with AWS monitoring and auditing services like CloudWatch (for metrics/logs) and CloudTrail (for API activity logging), giving visibility and compliance without custom instrumentation. Option C is correct because Bedrock provides a curated catalog of foundation models from multiple providers (e.g., Anthropic, AI21, Stability AI, Amazon Titan, Meta) through a single API, which self-managing on EC2 would require sourcing and integrating individually. Option D is correct because Bedrock is a fully managed service that handles model hosting, provisioning, and automatic scaling of inference capacity, removing the operational burden of managing EC2 instances, AMIs, autoscaling groups, and GPU capacity. Option B is not correct because Bedrock does not inherently reduce inter-region data transfer costs; those are governed by standard AWS data transfer pricing and are unrelated to the choice between Bedrock and self-managed EC2. Option E is not correct because self-managing foundation models on EC2 generally offers greater control over fine-tuning and customization, whereas Bedrock provides more constrained, managed customization options.

Answer

Lower data transfer costs between cloud regions.

Answer

Greater control over model fine-tuning and customization.

Question 3

A developer is building an application using Amazon Bedrock and needs to ensure that the model's responses do not include any toxic or harmful language. Which Bedrock feature should they configure?

Accepted Answer

Bedrock Guardrails. Bedrock Guardrails provide content filters, topic denial, PII detection, and other safety controls. Knowledge Bases, Agents, and Playground do not directly enforce content safety rules.

Answer

Bedrock Knowledge Bases

Answer

Bedrock Playground

Answer

Bedrock Agents

Question 4

A security analyst is reviewing CloudTrail logs for SageMaker API calls to identify which user executed a particular training job. The logs show assumed roles. In which CloudTrail event field can the analyst find the name of the user who assumed the role?

Accepted Answer

userIdentity.sessionContext.sessionIssuer.userName. When a user assumes an IAM role to perform SageMaker actions, the CloudTrail log records the assumed role's ARN in the `userIdentity.arn` field, but the original user's identity is preserved in the `userIdentity.sessionContext.sessionIssuer.userName` field. This field contains the name of the IAM user or role that initiated the `sts:AssumeRole` call, allowing the analyst to trace back to the actual user who assumed the role.

Answer

userIdentity.arn

Answer

eventName

Answer

requestParameters

Question 5

Which TWO actions help ensure fairness in an AI system deployed on AWS? (Select two.)

Accepted Answer

Train the model on a representative dataset. Option A (Train the model on a representative dataset) is correct because fairness in AI depends on the training data reflecting the demographics and conditions of the real-world population the model will serve; unrepresentative or skewed data leads to biased predictions. Option C (Use SageMaker Clarify to detect bias) is correct because SageMaker Clarify provides bias detection metrics (e.g., pre-training and post-training bias metrics such as Disparate Impact and Equal Opportunity Difference) that quantify and help mitigate bias in data and models. Option B (Enable AWS CloudTrail for audit) is not a fairness action; CloudTrail records API activity for security, compliance, and operational auditing, not for measuring or correcting model bias. Option D (Use a single validation set) is not a fairness measure and can actually increase evaluation variance; fairness requires representative and possibly multiple or cross-validated evaluation sets. Option E (Encrypt data at rest using AWS KMS) addresses data confidentiality and compliance, not fairness or bias in model outcomes.

Answer

Enable AWS CloudTrail for audit

Answer

Use a single validation set

Answer

Encrypt data at rest using AWS KMS

Question 6

Which TWO strategies can help reduce inference costs when using Amazon Bedrock? (Select TWO.)

Accepted Answer

Use provisioned throughput for high-volume, predictable workloads. Option C is correct because Provisioned Throughput on Amazon Bedrock lets you purchase dedicated model units at a fixed hourly rate for steady, high-volume, predictable inference traffic, which is typically cheaper than paying on-demand per-token prices once utilization is high. Option E is correct because choosing a smaller foundation model variant (for example, a mini or lite version of a model family) reduces the number of parameters and compute required per inference, directly lowering the per-token or per-request cost while often remaining sufficient for the task. Option A is not correct because temperature controls randomness of sampling, not the number of tokens generated, so raising it does not reduce cost and may even increase output variability. Option B is not correct because increasing max tokens allows longer responses, which increases the number of output tokens billed and therefore raises cost. Option D is not correct because Amazon ElastiCache is a caching service, not a native Bedrock cost-reduction strategy, and caching model responses is not one of the two intended Bedrock cost-optimization approaches in this scenario.

Answer

Use a higher temperature setting to generate fewer tokens

Answer

Increase the max tokens to allow longer responses

Answer

Cache frequently used responses in Amazon ElastiCache

Question 7

A startup uses Amazon Lex to build a chatbot for mental health support. They must ensure user conversations are private and not used for model improvement. Which AWS service can help anonymize text data before storage?

Accepted Answer

Amazon Comprehend. Amazon Comprehend offers a built-in feature called PII (Personally Identifiable Information) detection and redaction, which can automatically identify and mask sensitive data such as names, addresses, and health information in text. By using the `DetectPIIEntities` API with redaction, the startup can anonymize user conversations before storing them, ensuring compliance with privacy requirements and preventing data from being used for model improvement.

Answer

Amazon Textract

Answer

AWS Key Management Service (KMS)

Answer

Amazon Macie

Question 8

A company wants to evaluate the quality of a text generation model for a summarization task. They have reference summaries written by humans. Which automated metric compares the generated summary to the reference by measuring n-gram overlap?

Accepted Answer

ROUGE. ROUGE (Recall-Oriented Understudy for Gisting Evaluation) measures n-gram overlap between generated and reference summaries. BLEU is for translation, BERTScore uses embeddings, and perplexity measures language model confidence.

Answer

BLEU

Answer

Perplexity

Answer

BERTScore

Question 9

Which TWO actions can help reduce bias in a foundation model’s outputs? (Choose two.)

Accepted Answer

Fine-tune the model on a balanced, representative dataset. Option A is correct because fine-tuning on a balanced, representative dataset directly addresses bias by exposing the model to fair, diverse examples during training, which adjusts its learned weights and reduces skewed associations in its outputs. Option B is correct because careful prompt engineering with neutral wording avoids injecting biased framing or leading context into the input, so the model is less likely to amplify stereotypes or produce skewed responses. Option C is incorrect because restricting access to a subset of users is an access-control measure that does nothing to change the model's inherent biases. Option D is incorrect because increasing temperature only adds randomness to token sampling, which can make outputs more erratic rather than less biased. Option E is incorrect because a larger foundation model does not inherently reduce bias and may even reproduce or amplify biases present in its larger training corpus.

Answer

Restrict model access to a subset of users

Answer

Increase temperature to add randomness

Answer

Use a larger foundation model

Question 10

Which TWO actions are recommended for improving the factual accuracy of a foundation model's responses when using RAG?

Accepted Answer

Include relevant context from the knowledge base in the prompt. Option A is correct because RAG's core mechanism is retrieving relevant passages from the knowledge base and injecting them into the prompt as grounding context, which directly supplies the model with the facts it needs and reduces hallucination. Option C is correct because clear system-prompt instructions (for example, telling the model to answer only from the provided context and to say it doesn't know when the context is insufficient) constrain the model's behavior and improve factual fidelity. Option B is not recommended because max_tokens only caps response length and does not affect factual grounding. Option D is not recommended because a larger model does not guarantee factual accuracy and does not address retrieval quality. Option E is not recommended because raising temperature increases randomness and creativity, which typically worsens factual accuracy.

Answer

Increase the max_tokens parameter

Answer

Use the largest foundation model available

Answer

Increase the temperature parameter

Question 11

A healthcare organization uses an AI model to predict patient readmission risks. The model's predictions are used by doctors to allocate follow-up care. The organization wants to ensure compliance with responsible AI guidelines. Which practice best supports explainability?

Accepted Answer

Providing feature importance scores for each prediction. Feature importance scores directly address explainability by showing which input factors (e.g., age, lab results) most influenced each prediction. This allows doctors to understand and trust the model's reasoning, which is a core requirement of responsible AI guidelines for high-stakes healthcare decisions.

Answer

Ensuring the model's overall accuracy exceeds 95%

Answer

Using a black-box ensemble model that achieves highest accuracy

Answer

Automating decisions without human review to reduce bias

Question 12

A hospital uses an AI system to prioritize patients for organ transplant based on predicted survival rates. The system was trained on historical data that includes socioeconomic factors. A review reveals that the system systematically assigns lower priority to patients from lower-income neighborhoods, even when medical urgency is similar. The hospital's ethics board demands an immediate remedy. The data science team is small and must act quickly. What should the hospital do to address this fairness issue most effectively?

Accepted Answer

Retrain the model with only medically relevant features, after removing socioeconomic factors and correlated proxies. The bias stems from socioeconomic features and their correlated proxies leaking into the model, so the most effective remedy is to retrain using only medically relevant features after removing socioeconomic variables and any correlated proxies (e.g., ZIP code, insurance type). This addresses the root cause of the disparate impact rather than masking it. It is also feasible for a small team acting quickly, since it is a data and feature-engineering change rather than a full system rebuild.

Answer

Discontinue the AI system and have all prioritization done by a human committee

Answer

Apply a re-weighting penalty to boost priority for low-income patients

Answer

Use a different model type, such as a random forest instead of gradient boosting, on the same data

Question 13

A developer wants to use Amazon Bedrock to build a text summarization application. Which TWO of the following are required steps?

Accepted Answer

Request access to a foundation model in Amazon Bedrock. Option D is correct because Amazon Bedrock requires you to request and be granted access to a specific foundation model (for example, Amazon Titan Text or Anthropic Claude) in the target AWS Region before you can invoke it, since models are not enabled by default. Option E is correct because once access is granted, the application must call the model through the Bedrock runtime API (InvokeModel or Converse) using the AWS SDK or CLI, passing a prompt that contains the text to be summarized. Option A is incorrect because Bedrock is a fully managed, serverless service that exposes foundation models via API, so no SageMaker endpoint needs to be created or managed. Option B is incorrect because Bedrock provides pre-trained foundation models that can be used as-is or customized with techniques like fine-tuning or RAG, but training a model from scratch is not required. Option C is incorrect because Amazon EMR is not a prerequisite for Bedrock; any text preprocessing can be done in the application itself or with simpler services, and EMR is unnecessary for this scenario.

Answer

Create an Amazon SageMaker endpoint for hosting the model

Answer

Train a custom model from scratch on summarization data

Answer

Set up an Amazon EMR cluster to preprocess the text data

Question 14

A developer is using prompt engineering techniques to improve the performance of a text generation model on Amazon Bedrock. Which TWO techniques are examples of prompt engineering? (Select TWO.)

Accepted Answer

Few-shot prompting with example inputs and outputs. Few-shot prompting (B) is a core prompt engineering technique because it places several example input-output pairs directly in the prompt to steer the model's behavior without changing model weights. Zero-shot prompting (D) is likewise prompt engineering, as it crafts the instruction and task description so the model can perform the task with no examples provided. Both operate purely at the prompt/input level, which is the defining characteristic of prompt engineering on Amazon Bedrock. Fine-tuning (A) is excluded because it modifies model parameters through additional training rather than engineering the prompt. Adjusting temperature (C) is an inference parameter (decoding setting), not a prompt construction technique. Implementing a vector database for retrieval (E) is a Retrieval Augmented Generation architecture component, not prompt engineering itself.

Answer

Fine-tuning the model on domain-specific data

Answer

Adjusting the temperature parameter

Answer

Implementing a vector database for retrieval

Question 15

A company is optimizing costs for a Bedrock application that performs sentiment analysis on customer reviews. The workload is steady with occasional spikes. Which THREE strategies can help reduce costs without sacrificing accuracy? (Choose THREE)

Accepted Answer

Enable model caching to avoid reprocessing identical reviews. Option A is correct because enabling model caching (e.g., Bedrock prompt caching) stores results for identical review inputs, so repeated or duplicate reviews are served from cache instead of invoking the model again, directly cutting inference token costs while returning the same accurate output. Option B is correct because Bedrock batch inference processes large volumes of reviews asynchronously at a lower per-token price than on-demand invocation, and running these jobs during off-peak hours aligns with the steady workload with occasional spikes, reducing cost without changing model accuracy. Option D is correct because right-sizing to a smaller foundation model that still performs well on sentiment analysis lowers the per-token price and latency, and since accuracy is validated on the task, cost drops without sacrificing quality. Option C is not correct because fine-tuning a large model increases training and inference costs and is aimed at accuracy gains, not cost reduction, which the scenario does not need. Option E is not correct because provisioning throughput capacity (Provisioned Throughput) is a fixed hourly commitment suited to predictable high utilization; for a steady workload with only occasional spikes it can cost more than on-demand or batch usage rather than reducing costs.

Answer

Fine-tune a large model on the sentiment dataset for better accuracy

Answer

Provision enough throughput capacity to handle peak loads