Practice Exam AIF-C01: AWS Certified AI Practitioner

Question bank: 130 · timed mock exam: 90 min

Official exam page on the vendor's site (opens in a new tab) Official exam guide (opens in a new tab)

Exam guide: format, domains and how to prepare

Compare: AIF-C01 vs Google Generative AI Leader

Compare: CLF-C02 vs AIF-C01

Compare: AI-901 vs AIF-C01

  • Applications of Foundation Models 36 questions
  • Fundamentals of GenAI 32 questions
  • Fundamentals of AI and ML 26 questions
  • Guidelines for Responsible AI 18 questions
  • Security, Compliance, and Governance for AI Solutions 18 questions
Study mode (free) Exam mode (subscribers)

Free: 25 questions in study mode, all domains, no account.

Subscription: the 130 questions of this certification, the timed exam mode, and all other certifications too.

Subscribe · €9.99 / month

Secure card payment (Stripe) · No commitment · Cancel online

How to prepare for AWS Certified AI Practitioner with InfraPrep

The 130 questions cover 5 domains: Applications of Foundation Models, Fundamentals of GenAI, Fundamentals of AI and ML, Guidelines for Responsible AI, Security, Compliance, and Governance for AI Solutions. 25 of them are free, without an account.

Study mode

Answer at your pace: each answer is corrected straight away, with an explanation and the link to the documentation it relies on.

Exam mode

A timed exam of 90 minutes, with a question map to flag and come back to questions before you submit.

Results by domain

Your score is broken down by domain, so you know what to review first.

Original questions

Every question is written from the public documentation, never copied from the exam. You can contest an answer from study mode.

10 free sample questions

Question 1 Fundamentals of AI and ML

A product manager is preparing an internal glossary about artificial intelligence. One entry must describe the branch of machine learning that relies on neural networks with multiple layers and that is especially good at handling unstructured data such as images and text. Which term belongs in this entry?

  1. Generative AI
  2. Rule-based expert systems
  3. Deep learning
  4. Reinforcement learning
Show the answer and the explanation

Correct answer: C

AWS describes deep learning as a specialized subset of machine learning that uses neural networks with multiple layers, and notes that it is especially adept at unstructured data such as images and text.

Source: official documentation

Question 2 Fundamentals of GenAI

While reviewing the pricing of a text generation model, a finance analyst sees that usage is measured in tokens. The analyst asks what a token is. Which definition matches the AWS documentation?

  1. A sequence of characters that the model interprets or predicts as a single unit of meaning
  2. A vector of numerical values that represents the meaning of an entire document
  3. A value adjusted at inference time to change how varied the responses are
  4. A value that the provider updates to define how the model interprets its input
Show the answer and the explanation

Correct answer: A

AWS defines a token as a sequence of characters that a model can interpret or predict as a single unit of meaning, such as a word, part of a word, or a punctuation mark.

Source: official documentation

Question 3 Applications of Foundation Models

A European insurer uses Amazon Bedrock and wants to use cross-Region inference to absorb traffic peaks. A regulation requires that the processing of customer requests stays within the European Union. Which configuration should the insurer choose?

  1. A global cross-Region inference profile to get the largest pool of Regions
  2. Manual enablement of every commercial Region, then calling each Region in turn
  3. A geographic cross-Region inference profile tied to the EU geography
  4. A global cross-Region inference profile combined with provisioned throughput
Show the answer and the explanation

Correct answer: C

AWS recommends geographic cross-Region inference when there are data residency requirements and data processing must remain within specific geographic boundaries, such as the EU.

Source: official documentation

Question 4 Guidelines for Responsible AI

An AI governance committee maps its controls to the responsible AI dimensions defined by AWS. One control gives operators mechanisms to monitor the AI system and to steer or disable its behavior when necessary. Which dimension does this control address?

  1. Transparency
  2. Veracity and robustness
  3. Fairness
  4. Controllability
Show the answer and the explanation

Correct answer: D

AWS defines controllability as having mechanisms to monitor and steer AI system behavior.

Source: official documentation

Question 5 Security, Compliance, and Governance for AI Solutions

A company builds a customer chatbot on Amazon Bedrock. During a security review, the team asks who is responsible for preventing prompt injection vulnerabilities in the chatbot application. According to AWS documentation, who holds this responsibility?

  1. The foundation model provider that trained the model
  2. AWS and the customer share it equally for every prompt
  3. The customer, as part of secure application development
  4. AWS, because Amazon Bedrock hosts the foundation models
Show the answer and the explanation

Correct answer: C

AWS states that the responsibility for secure application development and preventing vulnerabilities like prompt injection lies with the customer, much like preventing SQL injection in database applications.

Source: official documentation

Question 6 Fundamentals of AI and ML

A retailer has several years of customer transaction histories. No customer has ever been tagged with a segment, but the marketing team wants to discover groups of customers with similar spending behavior. Which type of machine learning fits this requirement?

  1. Unsupervised learning with a clustering algorithm that groups similar customers
  2. Supervised learning with a classifier trained on existing segment labels
  3. Reinforcement learning with an agent that is rewarded for each purchase
  4. Supervised learning with a regression algorithm that predicts a spending amount
Show the answer and the explanation

Correct answer: A

The data has no labels and the goal is to group similar records; AWS lists clustering among the tasks addressed by unsupervised learning algorithms.

Source: official documentation

Question 7 Fundamentals of GenAI

A legal firm wants its internal search tool to return contracts that are similar in meaning to an employee's question, even when the question and the contract use different words. Which technique addresses this need?

  1. Fine-tune a foundation model with labeled pairs of questions and contract excerpts
  2. Convert the contracts and the question into vector embeddings and compare the vectors
  3. Lower the temperature so that the model always selects higher-probability tokens
  4. Split each contract into tokens and return contracts that share identical token sequences
Show the answer and the explanation

Correct answer: B

Embeddings are series of numbers that allow texts to be compared mathematically, and they are used to determine the semantic similarity between queries and the text of the data sources.

Source: official documentation

Question 8 Applications of Foundation Models

An online retailer wants a generative AI assistant that answers questions about its product catalog. The catalog changes every day, and the retailer wants accurate answers based on the latest data while keeping costs low. Which design best meets these requirements?

  1. Connect the catalog to a knowledge base and use retrieval augmented generation
  2. Fine-tune the foundation model every night on the updated catalog
  3. Distill a smaller student model from a larger teacher model every week
  4. Increase the top K value so that the model considers more candidate tokens
Show the answer and the explanation

Correct answer: A

A knowledge base retrieves current information from the data source at query time and removes the need to continually train the model to use private data, which keeps answers current and costs lower.

Source: official documentation

Question 9 Guidelines for Responsible AI

A bank is launching a generative AI assistant for its customers. The assistant must refuse to discuss illegal investment advice, however the user phrases the request. Which Amazon Bedrock Guardrails policy is designed for this requirement?

  1. A denied topic that describes illegal investment advice
  2. A sensitive information filter that masks personal data
  3. A word filter that lists specific banned words and phrases
  4. A contextual grounding check against the retrieved sources
Show the answer and the explanation

Correct answer: A

Denied topics let you define topics to avoid in a generative AI application; AWS gives the example of a banking assistant that avoids topics related to illegal investment advice.

Source: official documentation

Question 10 Security, Compliance, and Governance for AI Solutions

A healthcare company runs Amazon Bedrock batch inference and model customization jobs. Its security policy states that the data used by these jobs must not be available over the internet and must travel over a private connection. What should the company configure?

  1. An IAM policy that allows only the bedrock:InvokeModel action
  2. A VPC with an interface VPC endpoint created with AWS PrivateLink
  3. Amazon Macie discovery jobs on the S3 buckets that hold the data
  4. Customer managed AWS KMS keys for the job output only
Show the answer and the explanation

Correct answer: B

AWS recommends configuring a VPC so that data isn't available over the internet and creating an interface VPC endpoint with AWS PrivateLink to establish a private connection to the data.

Source: official documentation

Frequently asked questions

Are there free questions, and do I need an account?

Yes. 25 questions of this certification are free in study mode, across all its domains, without a subscription and without registration. The account is created when you subscribe, at checkout.

How does the subscription work, and can I cancel?

One monthly subscription covers all certifications and the timed practice exams. There is no minimum period. You cancel online, at any time, with the "Cancel your contract" function in My account; cancellation takes effect at the end of the month already paid, and no further payment is taken.

Where do the questions come from?

Every question is original, written from the vendor's public documentation. Each answer comes with an explanation and a link to the documentation page it relies on. InfraPrep is not affiliated with the vendors, and certification names belong to their owners.

In which language are the questions?

The questions are written in English, also on the French version of the site.

Is the catalog updated?

Yes, the catalog changes over time: new certifications and new questions are added.