10 free Google Generative AI Leader practice questions with the correct answer and a full explanation for each, taken from the CertStash pack of 75 questions. Work through them, then open each answer to check your reasoning.
Get all 75 questions (US$39) · Download these 10 as a PDF
Question 1
What is the definition of generative AI?
Show answer and explanation
Correct answer: B. A type of artificial intelligence that can create new content and ideas, including text, images, music, and code.
Generative AI is specifically designed to create new content across multiple formats including text, images, music, and code. This distinguishes it from other AI approaches that focus on prediction, pattern recognition through neural networks, or autonomous learning. The defining characteristic of generative AI is its capability to produce novel outputs rather than simply analyze or classify existing data.
Why the other options are wrong
- A. This describes linear regression, a predictive modeling technique, not generative AI.
- C. This describes artificial neural networks, which are a computational architecture used in various AI types, not specifically generative AI.
- D. This describes general machine learning with deep learning, which is broader than generative AI and doesn't capture the content creation aspect.
Question 2
A company wants to use an AI agent to automate some tasks. They want everyone to understand the different functions of an AI agent.
What is the function of an AI agent in the context of gen AI?
Show answer and explanation
Correct answer: D. To analyze situations, use multiple tools, and make informed decisions without requiring constant human input.
An AI agent in the context of generative AI serves as an autonomous decision-maker that analyzes situations, leverages multiple available tools, and makes informed decisions independently without requiring constant human intervention. This is the core function that distinguishes an agent from other components in a gen AI system, it acts as the intelligent executor that orchestrates the use of various resources to accomplish objectives.
Why the other options are wrong
- A. This describes computational infrastructure, not the function of an AI agent.
- B. This describes data storage and management systems, which are supporting infrastructure rather than the agent itself.
- C. This describes a user interface layer, which is separate from the agent's core function of autonomous decision-making.
Question 3
What is the definition of an AI agent?
Show answer and explanation
Correct answer: A. An application that learns how to achieve a goal based on inputs and tools available to it.
An AI agent is fundamentally an application that possesses the ability to learn and adapt based on available inputs and tools to achieve defined goals. This learning capability is essential, it distinguishes an agent from static systems and emphasizes the dynamic, goal-oriented nature of agents that autonomously improve their performance over time.
Why the other options are wrong
- B. This describes a human role or oversight function, not an AI agent itself.
- C. This describes a user interface component, not the definition of an AI agent.
- D. This describes a static information provider, lacking the learning and goa-chievement capabilities that define an AI agent.
Question 4
A company is developing an AI character for a video game. The AI character needs to learn how to navigate a complex environment and make decisions to achieve certain objectives within the game.
When the AI takes actions that lead to positive outcomes like finding a reward or overcoming an obstacle, it receives a positive score.
When it takes actions that lead to negative outcomes, like hitting a wall or losing progress, it receives a negative score. Through this process of trial and error, the AI gradually improves the character’s ability to play the game effectively.
What machine learning should the company use?
Show answer and explanation
Correct answer: B. Reinforcement learning
Reinforcement learning is the appropriate machine learning approach for this scenario because the AI character learns through a reward-based feedback system. The character receives positive scores for successful actions and negative scores for unsuccessful ones, and through this iterative trial-and-error process, it gradually optimizes its behavior to maximize rewards. This defines the core mechanism of reinforcement learning.
Why the other options are wrong
- A. Unsupervised learning finds patterns in unlabeled data without explicit reward signals, which doesn't match this scenario's scored feedback system.
- C. Supervised learning requires pre-labeled training data with correct answers, but this scenario involves learning through environmental feedback and scores.
- D. Deep learning is a technique that can be combined with other approaches but doesn't specifically describe the reward-based learning mechanism being used here.
Question 5
A research team has collected a large dataset of sensor readings from various industrial machines. This dataset includes measurements like temperature, pressure vibration levels and electrical current recorded at regular intervals. The team has not yet assigned any labels or categories to these readings and wants to identify potential anomalies, malfunctions or natural groupings of machine behavior based on the sensor data alone.
What type of machine learning should they use?
Show answer and explanation
Correct answer: D. Unsupervised learning
Unsupervised learning is the correct approach because the dataset has no predefined labels or categories, and the goal is to discover natural patterns and groupings within the sensor data. The team aims to identify anomalies and cluster machine behaviors without guidance from labeled examples, which is precisely what unsupervised learning techniques like clustering and anomaly detection accomplish.
Why the other options are wrong
- A. Deep learning is a technique that could be applied within supervised or unsupervised contexts, but it doesn't specifically address the unlabeled nature of this dataset.
- B. Reinforcement learning requires a reward signal and agent-environment interaction, which is not part of this static sensor data analysis scenario.
- C. Supervised learning requires labeled data with known correct answers, but this dataset contains no labels or predefined categories.
Question 6
A highly regulated financial institution wants to use Gemini as the core decision engine for a loan approval system that will deterministically approve or reject loan applications based on a strict set of predefined criteria.
Why is this an inappropriate use case for Gemini?
Show answer and explanation
Correct answer: B. Gemini’s designed for flexible content generation and inference, not rigid rul-ased decisions.
Gemini is designed for flexible, generative tasks that produce varied content and contextual responses. A loan approval system requires deterministic, rule-based decisio-aking where the same input must always produce the same output for regulatory compliance. Gemini's stochastic nature and tendency to generate flexible inferences make it fundamentally unsuitable for rigid, rule-based systems where consistency and reproducibility are legally mandated.
Why the other options are wrong
- A. Cost and complexity are operational considerations, not fundamental architectural reasons why Gemini is inappropriate for deterministic rule-based decisions.
- C. Gemini can process and reason about structured numerical data; this is not the core limitation for this use case.
- D. Database integration is a technical implementation detail that doesn't address why a generative model is inappropriate for rigid rule-based approval logic.
Question 7
What is an example of unsupervised machine learning?
Show answer and explanation
Correct answer: C. Analyzing customer purchase patterns to identify natural groupings.
Analyzing customer purchase patterns to identify natural groupings is an example of unsupervised learning because it involves clustering or segmenting customers based on their purchasing behavior without predefined labels or categories. The system discovers patterns and relationships in the data independently, without being trained on labeled examples of what constitutes a specific customer segment.
Why the other options are wrong
- A. Predicting renewal based on past status data is supervised learning, as it uses labeled historical outcomes to predict future behavior.
- B. Training on labeled product images is supervised learning, as the categories and correct answers are predefined.
- D. Forecasting sales using historical data is supervised learning, as it uses labeled historical examples to predict a specific outcome.
Question 8
A user asks a generative AI model about the scientific accuracy of a popular science fiction movie. The model confidently states that humans can indeed travel faster than light, referencing specific but entirely fictional theories and providing made-up explanations of how this is achieved according to the movie's "established science." The model presents this information as factual, without indicating that it originates from a fictional work.
What type of model limitation is this?
Show answer and explanation
Correct answer: B. Hallucination
Hallucination is the model limitation being demonstrated. The model is confidently generating false information that appears plausible but is entirely fabricated, inventing fictional theories and presenting them as established facts without acknowledging their fictional origin. This is the defining characteristic of hallucination: the model produces convincing but false outputs that it presents with unwarranted confidence.
Why the other options are wrong
- A. Bias refers to systematic errors or prejudices in model behavior, not the generation of false information.
- C. Knowledge cutoff refers to the model's training data having a temporal limit, not the generation of entirely fabricated theories.
- D. Data dependency describes how model performance varies with training data quality, not the phenomenon of generating false confident statements.
Question 9
What does a diffusion model do?
Show answer and explanation
Correct answer: A. Generates high-quality content by refining noise into structured data.
Diffusion models work by iteratively refining noisy data into structured, high-quality outputs through a denoising process. They start with random noise and progressively remove that noise to generate coherent content such as images, text, or other data types. This reverse diffusion process is the core mechanism that enables generative capabilities.
Why the other options are wrong
- B. Data storage and management are infrastructure concerns unrelated to what diffusion models actually do.
- C. Predictive analytics and trend forecasting describe discriminative or forecasting models, not generative diffusion models.
- D. Business process optimization is an operational outcome, not the fundamental function of a diffusion model.
Question 10
An organization is collecting data to train a generative AI model for customer service. They want to ensure security throughout the ML lifecycle.
What is a critical consideration at this stage?
Show answer and explanation
Correct answer: D. Implementing access controls and protecting sensitive information within the training data.
During the data collection and training stage of the ML lifecycle, protecting sensitive information in the training data through access controls is critical. Training data often contains personal or confidential information, and without proper protections, this data can be exposed, compromised, or misused. This foundational security measure must be established early to maintain data confidentiality throughout the entire ML pipeline.
Why the other options are wrong
- A. Software patching applies to deployment and maintenance phases, not the training data collection stage.
- B. Monitoring model performance for unexpected outputs is important post-deployment, not during initial data collection.
- C. Ethical guidelines for responses are relevant during model design and deployment, but data security during collection is the more critical foundational step.
That was 10 of 75.
The full Google Generative AI Leader pack has all 75 questions, each with the answer, the explanation and why the other options are wrong, plus a questions-only copy for timed runs. US$39, paid once, with free monthly updates and a pass-or-your-money-back guarantee.