AWS AIF-C01: AWS Certified AI Practitioner

Last checked October 22, 2026 against the official AWS exam page.

1. Question 1Fundamentals of AI and ML

   A retailer wants to predict next month's sales for each store using historical numeric data. Which type of machine learning problem is this?

   1. ARegression
   2. BAnomaly detection with no labels
   3. CClustering
   4. DReinforcement learning
   Show answer and explanationHide answer

   Correct answer: A

   Predicting a continuous numeric value from labeled historical data is regression. Clustering groups unlabeled data, reinforcement learning learns from rewards, and anomaly detection flags unusual points rather than predicting a value.

2. Question 2Applications of Foundation Models

   A company wants a model to answer questions using its internal policy documents, with answers that reflect the latest versions without retraining. Which approach fits best?

   1. AIncreasing the temperature to 1.5
   2. BRetrieval Augmented Generation (RAG) with a knowledge base over the documents
   3. CFine-tuning the model once a year on old policies
   4. DAsking the model to memorize the documents in its prompt each time without retrieval
   Show answer and explanationHide answer

   Correct answer: B

   RAG retrieves current passages from the document store and adds them to the prompt, so answers follow updated documents without retraining. Annual fine-tuning lags behind changes, high temperature increases randomness, and manual prompting does not scale.

3. Question 3Fundamentals of GenAI

   What is a foundation model in generative AI?

   1. AA model that works only on a single predefined task
   2. BA rule-based system with fixed if-then statements
   3. CA large model pre-trained on broad data that can be adapted to many downstream tasks
   4. DA dataset used to store embeddings
   Show answer and explanationHide answer

   Correct answer: C

   Foundation models are large, pre-trained models that can be adapted or prompted for many tasks. Single-task models, datasets and rule-based systems do not match this definition.

4. Question 4Security, Compliance, and Governance for AI Solutions

   A company wants to ensure that only approved users can invoke a generative AI model in Amazon Bedrock. Which control should be used?

   1. ASharing one long-term IAM access key across the developer team for service calls
   2. BDisabling AWS CloudTrail logging in the account so that Bedrock usage stays private
   3. CMaking the Bedrock runtime endpoint public with no authentication
   4. DIAM policies that grant bedrock invoke permissions only to approved roles
   Show answer and explanationHide answer

   Correct answer: D

   IAM policies restrict which principals can invoke Bedrock models, which enforces least privilege. Public endpoints, shared keys and disabled audit trails weaken security and governance.

5. Question 5Guidelines for Responsible AI

   A lending company's model approves applications at very different rates for two demographic groups with similar credit profiles. Which concept describes this issue?

   1. ABias and fairness problems in model outcomes
   2. BModel drift from shifts in incoming data
   3. CLatency in inference responses during peak
   4. DData residency restrictions by region
   Show answer and explanationHide answer

   Correct answer: A

   Different outcomes for similar profiles across groups indicates bias or unfairness. Drift refers to changing data over time, latency is a performance measure, and residency is a data location requirement.