All certifications
AIAmazon Web ServicesAIF-C01
AWS Certified AI Practitioner
Exam overview, official domain weights and free sample questions. Subscribers get 229 practice questions, 112 flashcards, 3 labs, a study guide and video lessons for this exam.
Exam at a glance
- Exam code
- AIF-C01
- Questions
- 65 (50 scored + 15 unscored)
- Time
- 90 minutes
- Question types
- multiple choice, multiple response, ordering, matching
- Passing score
- 700 (scaled 100–1,000)
- Exam fee
- 100 USD
- Delivery
- Pearson VUE test center, online proctored
- Prerequisites
- None. Recommended: up to 6 months of exposure to AI/ML technologies on AWS
- Renewal
- Valid 3 years. Recertify by passing the latest version, or by earning AWS Certified Machine Learning Engineer – Associate
- Version notes
- No replacement announced on the official page. Italian and German versions retire after October 15, 2026
From the official exam page (checked 2026-11-09). Always confirm details with the vendor: official exam page, official objectives.
Domains and weights
Fundamentals of AI and ML20%
- 1.1 Explain basic AI concepts and terminologies
- 1.2 Identify practical use cases for AI
- 1.3 Describe the AI/ML development lifecycle
Fundamentals of Generative AI24%
- 2.1 Explain the basic concepts of generative AI (GenAI)
- 2.2 Understand the capabilities and limitations of GenAI for solving business problems
- 2.3 Describe AWS infrastructure and technologies for building GenAI applications
Applications of Foundation Models28%
- 3.1 Describe design considerations for applications that use foundation models (FMs)
- 3.2 Choose effective prompt engineering techniques
- 3.3 Describe the training and fine-tuning process for FMs
- 3.4 Describe methods to evaluate FM performance
Guidelines for Responsible AI14%
- 4.1 Explain the development of AI systems that are responsible
- 4.2 Recognize the importance of transparent and explainable models
Security, Compliance, and Governance for AI Solutions14%
- 5.1 Explain methods to secure AI systems
- 5.2 Recognize governance and compliance regulations for AI systems
Free sample questions
Original questions in the style of the exam. Pick an answer to see the explanation.
Sample question 1 · Fundamentals of AI and ML
A retailer has five years of labeled transactions marked 'fraud' or 'legitimate' and wants to flag new transactions automatically. Which type of machine learning fits best?
Correct — “Supervised learning (binary classification)”: Labeled historical outcomes plus a yes/no target is the textbook case for supervised classification. Incorrect — “Dimensionality reduction”: PCA-style reduction compresses features; it does not by itself predict fraud. Incorrect — “Unsupervised clustering of transactions into similar groups”: Clustering ignores the labels the retailer already has and does not predict a named class. Incorrect — “Reinforcement learning”: RL learns from rewards through interaction with an environment, not from a labeled historical dataset.
Sample question 2 · Fundamentals of Generative AI
What is a token in the context of large language models?
Incorrect — “An authentication credential that the application sends to the model API”: That is an API key or access token, a different meaning of the word. Correct — “A unit of text, such as a word or part of a word, that the model processes”: LLMs split text into tokens; limits and pricing are counted in tokens. Incorrect — “A single training example”: Training examples contain many tokens. Incorrect — “A probability score assigned to each candidate answer”: Models compute probabilities over tokens, but a token is the unit of text itself.
Sample question 3 · Applications of Foundation Models
A legal-research firm needs answers grounded in its own 200,000 internal memos, which change weekly. Which approach is most cost-effective and keeps answers current?
Incorrect — “Increase the temperature”: Temperature does not add knowledge. Incorrect — “Fine-tune the foundation model every day on the full set of 200,000 memos”: Daily fine-tuning is expensive and still risks stale or hallucinated facts. Correct — “Retrieval Augmented Generation (RAG) using a knowledge base”: RAG retrieves current documents at query time, so no retraining is needed when memos change. Incorrect — “Pre-train a new foundation model from scratch”: Pre-training is extremely costly and would be stale within a week.
Unlock every question, lab and guide
Full access to all 20 certifications for $4.99/month. Cancel anytime.