🎓 AI Fundamentals Course: Final Proficiency Exam
Question 1 (Multiple Choice – Module 1)
What is the underlying foundational technology that powers user-facing conversational interfaces like ChatGPT, Claude, and Copilot?
- A) An autonomous AI Agent
- B) A Large Language Model (LLM)
- C) A static keyword database
- D) A cloud-based web scraper
Correct Answer: B
Explanation: A Large Language Model (LLM) is the core engine trained on billions of pages of text to calculate word probabilities. Tools like ChatGPT are applications built on top of these models.
Explanation: A Large Language Model (LLM) is the core engine trained on billions of pages of text to calculate word probabilities. Tools like ChatGPT are applications built on top of these models.
Question 2 (True / False – Module 1)
Statement: AI systems are inherently sentient, meaning they possess subjective feelings, personal beliefs, and an actual human-like understanding of the text they process.
Correct Answer: False
Explanation: AI does not have consciousness, feelings, or lived experiences. It mathematically processes structural text patterns based on probabilities without experiencing real understanding.
Explanation: AI does not have consciousness, feelings, or lived experiences. It mathematically processes structural text patterns based on probabilities without experiencing real understanding.
Question 3 (Fill in the Blank – Module 1)
When a Large Language Model generates a response that sounds completely realistic and confident but is factually incorrect or entirely fabricated, this phenomenon is called a ________.
Correct Answer: hallucination / hallucinations
Explanation: Hallucinations occur when an LLM prioritizes linguistic probability over factual accuracy, generating plausible-sounding falsehoods.
Explanation: Hallucinations occur when an LLM prioritizes linguistic probability over factual accuracy, generating plausible-sounding falsehoods.
Question 4 (Multiple Choice – Module 2)
Which core component of a prompt provides the background data, situational scenario, or specific domain information necessary for the AI to tailor its response effectively?
- A) The Clear Instruction
- B) The Explicit Constraint
- C) The Sufficient Context
- D) The Output Format
Correct Answer: C
Explanation: Context serves as the background framing or scenario that grounds the AI’s generation, ensuring the output is highly relevant to the user’s specific situation.
Explanation: Context serves as the background framing or scenario that grounds the AI’s generation, ensuring the output is highly relevant to the user’s specific situation.
Question 5 (True / False – Module 2)
Statement: Adding polite filler phrases such as “If it’s not too much trouble, could you please kindly draft…” helps the AI process commands faster because it mimics polite human relationships.
Correct Answer: False
Explanation: Conversational filler occupies valuable space in the AI’s processing window. High-quality prompts are direct, stating exactly what is expected or required without conversational padding.
Explanation: Conversational filler occupies valuable space in the AI’s processing window. High-quality prompts are direct, stating exactly what is expected or required without conversational padding.
Question 6 (Multiple Choice – Module 2)
An AI reads text sequentially and evaluates prompt structure deliberately. Which areas of a prompt carry the most structural weight and are followed most reliably by the model?
- A) The exact middle paragraphs of the prompt
- B) Phrases hidden within brackets inside background data
- C) The very beginning or the absolute end of the prompt
- D) Sentences containing polite conversational filler
Correct Answer: C
Explanation: Due to the sequential nature of attention mechanisms in LLMs, instructions placed at the absolute start or the very end of a prompt are prioritized much more effectively than those buried in the middle.
Explanation: Due to the sequential nature of attention mechanisms in LLMs, instructions placed at the absolute start or the very end of a prompt are prioritized much more effectively than those buried in the middle.
Question 7 (Fill in the Blank – Module 2)
Burying your primary request deep inside a massive paragraph of background text is a beginner mistake known as hiding the core ________.
Correct Answer: task / instruction / requirement
Explanation: Hiding the core task confuses the model’s focus, as it struggles to distinguish between foundational context and the explicit action it is expected to perform.
Explanation: Hiding the core task confuses the model’s focus, as it struggles to distinguish between foundational context and the explicit action it is expected to perform.
Question 8 (Multiple Choice – Module 3)
What does the acronym RGCCO stand for in advanced prompt engineering?
- A) Rules, Goals, Constraints, Context, Outcome
- B) Role, Goal, Context, Constraints, Output
- C) Research, Guidelines, Category, Criteria, Objective
- D) Role, Guidance, Core, Clarity, Operational
Correct Answer: B
Explanation: The RGCCO framework represents the five essential pillars of a bulletproof prompt: Role (persona), Goal (primary task), Context (background), Constraints (boundaries), and Output (formatting).
Explanation: The RGCCO framework represents the five essential pillars of a bulletproof prompt: Role (persona), Goal (primary task), Context (background), Constraints (boundaries), and Output (formatting).
Question 9 (True / False – Module 3)
Statement: Within the RGCCO formula, the “Role” and the “Context” function in complete isolation, meaning the background scenario should never influence or change what constraints you write.
Correct Answer: False
Explanation: RGCCO components are interdependent. For example, if your Context defines your target audience as “absolute beginners,” that specific context dictates constraints against using technical jargon.
Explanation: RGCCO components are interdependent. For example, if your Context defines your target audience as “absolute beginners,” that specific context dictates constraints against using technical jargon.
Question 10 (Multiple Choice – Module 3)
Review this snippet from a copywriting prompt: “Do not use generic marketing buzzwords like ‘synergy,’ ‘disruptive,’ or ‘paradigm shift’.” Which RGCCO pillar does this statement belong to?
- A) Goal
- B) Context
- C) Role
- D) Constraints
Correct Answer: D
Explanation: Negative rules that explicitly state what the AI must avoid, prohibit, or restrict fall directly under the Constraints category.
Explanation: Negative rules that explicitly state what the AI must avoid, prohibit, or restrict fall directly under the Constraints category.
Question 11 (Fill in the Blank – Module 3)
In the RGCCO framework, the component that dictates whether your final answer should be delivered as a markdown table, a bulleted list, or an email template is the ________ parameter.
Correct Answer: Output / O
Explanation: The Output parameter defines the physical layout, structural packaging, and presentation style of the AI’s final response.
Explanation: The Output parameter defines the physical layout, structural packaging, and presentation style of the AI’s final response.
Question 12 (Multiple Choice – Module 4)
You need to generate an Instagram caption promoting a local gym. Which of the following prompts represents the most optimized application of the RGCCO framework?
- A) Write an Instagram post about fitness and working out at our local gym.
- B) Act as a fitness social media manager. Write an Instagram caption promoting our local gym. Target busy working parents with a hook in the first sentence. Avoid cheesy cliches. Present as a scannable bullet list.
- C) Can you please write a quick post for Instagram? I need it to be about my gym, and if you don’t mind, make it sound really energetic and cool for people who want to lose weight.
- D) Create an Instagram caption about a local gym. The gym is open 24/7, has free childcare, and offers group classes.
Correct Answer: B
Explanation: Option B flawlessly covers all RGCCO pillars: Role (social media manager), Goal (write caption), Context (working parents/local gym), Constraints (avoid cliches/hook first), and Output (scannable bullet list).
Explanation: Option B flawlessly covers all RGCCO pillars: Role (social media manager), Goal (write caption), Context (working parents/local gym), Constraints (avoid cliches/hook first), and Output (scannable bullet list).
Question 13 (True / False – Module 4)
Statement: When prompting an AI to write educational material for beginners, it is safe to assume the AI will automatically know to avoid advanced, industry-specific acronyms even if you do not specify any constraints.
Correct Answer: False
Explanation: AI cannot read your mind. If you do not explicitly place a constraint banning advanced jargon, the AI will default to its standard text distributions, which frequently include complex terminology.
Explanation: AI cannot read your mind. If you do not explicitly place a constraint banning advanced jargon, the AI will default to its standard text distributions, which frequently include complex terminology.
Question 14 (Fill in the Blank – Module 4)
When writing an effective prompt, utilizing an explicit action verb such as “Analyze,” “Summarize,” or “Execute” is the best way to define a clear ________.
Correct Answer: Goal / G / task / instruction
Explanation: The Goal must always be driven by a strong, unambiguous action verb so the software understands the precise operation it needs to perform.
Explanation: The Goal must always be driven by a strong, unambiguous action verb so the software understands the precise operation it needs to perform.
Question 15 (Multiple Choice – Module 5)
What happens mathematically inside an LLM if your prompt accidentally contains heavily conflicting words or contradictory instructions?
- A) The AI pauses execution and prompts the user for clarification.
- B) The AI automatically ignores the second instruction and executes the first perfectly.
- C) The AI treats text as an associative map, resulting in a muddled, corrupted, or highly inaccurate response.
- D) The AI overrides your text to execute standard ethical guidelines instead.
Correct Answer: C
Explanation: LLMs process text as a network of structural associations. Forcing the model to associate contradictory concepts pulls the vector mathematics in opposing directions, fracturing the quality of the logic.
Explanation: LLMs process text as a network of structural associations. Forcing the model to associate contradictory concepts pulls the vector mathematics in opposing directions, fracturing the quality of the logic.
Question 16 (True / False – Module 5)
Statement: Advanced prompt optimization requires specifying clear criteria for how the AI should score or evaluate information, particularly when asking it to perform objective tasks like a financial variance audit.
Correct Answer: True
Explanation: Providing clear analytical parameters and benchmarking criteria prevents the AI from relying on subjective or generic logic when evaluating complex datasets.
Explanation: Providing clear analytical parameters and benchmarking criteria prevents the AI from relying on subjective or generic logic when evaluating complex datasets.
Question 17 (Fill in the Blank – Module 5)
To prevent an AI from outputting generic, uninspired industry clichés, prompt engineers use ________ constraints to explicitly list forbidden words or phrases.
Correct Answer: negative / exclusionary / prohibition
Explanation: Negative constraints act as direct blocklists that steer the AI away from predictable linguistic paths, forcing it to generate more original or precise copy.
Explanation: Negative constraints act as direct blocklists that steer the AI away from predictable linguistic paths, forcing it to generate more original or precise copy.
Question 18 (Multiple Choice – Module 6)
What is the core operational feature that defines an autonomous AI Agent?
- A) It requires a human to type a new command at every single step of its task.
- B) It is given a broad, overarching destination and independently plans steps, uses digital tools, and checks its own work.
- C) It operates purely inside a static text chat window and cannot view external files.
- D) It replaces the Large Language Model entirely with manual hardcoded rules.
Correct Answer: B
Explanation: Unlike traditional chatbots that stop and wait for human input, an AI Agent uses an LLM as a logical processing engine to autonomously plan, execute tool loops, and evaluate its own progress until a broad goal is met.
Explanation: Unlike traditional chatbots that stop and wait for human input, an AI Agent uses an LLM as a logical processing engine to autonomously plan, execute tool loops, and evaluate its own progress until a broad goal is met.
Question 19 (True / False – Module 6)
Statement: According to the standard AI Agent Flow, an agent operates via a continuous loop where it executes a tool, reviews the output of that tool against the objective, and adjusts its subsequent actions independently.
Correct Answer: True
Explanation: The agent architecture relies on a continuous feedback loop: [Goal] ➔ [Plan] ➔ [Execute Tool] ➔ [Check Work]. It actively monitors its performance until the target state is reached.
Explanation: The agent architecture relies on a continuous feedback loop: [Goal] ➔ [Plan] ➔ [Execute Tool] ➔ [Check Work]. It actively monitors its performance until the target state is reached.
Question 20 (Multiple Choice – Module 6)
Which of the following scenarios best demonstrates a real-world corporate application of a specialized AI Agent workflow?
- A) A text window that answers basic employee questions about the company holiday policy when asked.
- B) A software tool that requires an HR manager to manually download, review, and score every single incoming resume.
- C) A system that automatically ingests a job application, verifies credentials via a database, deploys a screening test, evaluates results, and schedules an interview without manual human intervention.
- D) A basic digital calendar that sends a generic notification when an appointment is manually created.
Correct Answer: C
Explanation: Option C describes an end-to-end autonomous loop where the system assesses its environment, triggers multiple software tools independently, evaluates data, and achieves a final goal without manual, step-by-step human management.
Explanation: Option C describes an end-to-end autonomous loop where the system assesses its environment, triggers multiple software tools independently, evaluates data, and achieves a final goal without manual, step-by-step human management.
Question 21 (Fill in the Blank – Module 6)
While a standard chatbot is locked inside a back-and-forth conversational loop, an AI agent is capable of interacting with its environment by opening and using external digital ________.
Correct Answer: tools / applications / software / APIs
Explanation: AI agents extend their capabilities beyond simple text generation by leveraging specialized digital tools (like web browsers, calculators, databases, and email clients) to execute complex real-world workflows.
Explanation: AI agents extend their capabilities beyond simple text generation by leveraging specialized digital tools (like web browsers, calculators, databases, and email clients) to execute complex real-world workflows.
Question 22 (Multiple Choice – Module 7)
You are conducting hands-on prompt testing and notice that the AI keeps generating summaries that are far too long and detailed. According to prompt optimization best practices, how should you fix this prompt?
- A) Add a polite closing phrase asking the AI to be brief if it has time.
- B) Inject a strict, numeric negative constraint at the absolute end of your prompt, such as: “Constraint: Your summary must be under 150 words total.”
- C) Write a completely separate sentence explaining the entire history of why you need a short summary.
- D) Keep the prompt exactly the same but run it multiple times hoping it fixes itself.
Correct Answer: B
Explanation: Adding a strict, numeric constraint and placing it at the end of the prompt utilizes proper structural prioritization to force compliance over the output’s length.
Explanation: Adding a strict, numeric constraint and placing it at the end of the prompt utilizes proper structural prioritization to force compliance over the output’s length.
Question 23 (True / False – Module 7)
Statement: When practicing prompt structure, running an overly brief prompt like “Write a report on marketing” is considered an effective engineering benchmark because it lets the AI choose the best path forward.
Correct Answer: False
Explanation: Overly brief prompts force the AI to guess your intent, audience, and formatting. This inevitably results in highly generic, unhelpful, and artificial-looking text that requires heavy editing.
Explanation: Overly brief prompts force the AI to guess your intent, audience, and formatting. This inevitably results in highly generic, unhelpful, and artificial-looking text that requires heavy editing.
Question 24 (Multiple Choice – Module 7)
What is the primary benefit of transforming your daily digital workflows into the RGCCO prompt structure?
- A) It makes the AI generate responses that look exactly like human-written code scripts.
- B) It completely guarantees that the AI will never experience a hallucination under any circumstances.
- C) It drastically minimizes AI guessing, ensuring highly predictable, tailored, and actionable outputs on the very first try.
- D) It allows the AI to run autonomously in the background as a full agent without any initial user input.
Correct Answer: C
Explanation: The true power of the RGCCO framework is the systematic elimination of contextual and stylistic ambiguity, unlocking clean, highly utility-dense answers immediately and ending the cycle of endless revisions.
Explanation: The true power of the RGCCO framework is the systematic elimination of contextual and stylistic ambiguity, unlocking clean, highly utility-dense answers immediately and ending the cycle of endless revisions.
Question 25 (Fill in the Blank – Module 7)
If your input prompt is unstructured, vague, and full of contradictory terminology, the resulting AI generation will reflect those flaws. This fundamental data processing rule is commonly known as “Poor input leads directly to poor ________.”
Correct Answer: output / generation / results / response
Explanation: Because LLMs are predictive engines mapped to user associations, the precision of the output is completely bound to the explicit structure and clarity of the input text provided.
Explanation: Because LLMs are predictive engines mapped to user associations, the precision of the output is completely bound to the explicit structure and clarity of the input text provided.