AI literacy isn’t just about knowing what to ask—it’s about crafting prompts that spark curiosity, deepen understanding, and support ethical, critical exploration.
In this module, you learned how an AI prompt differs from a traditional Google or keyword search. While search engines retrieve links based on keywords, large language models (LLMs) generate text by predicting the most probable next word based on patterns in their training data. Some newer systems (like Copilot and Perplexity) use hybrid approaches, such as Retrieval-Augmented Generation (RAG), that combine LLMs with real-time access to external sources, helping to ground their responses in more current or verifiable information.
However, unlike humans, AI models cannot truly understanding context or verify facts. They generate language based on probability, not evidence. Because of this, the quality of an AI's response depends heavily on how the prompt is written—not just the words you use, but also the clarity, structure, and context you provide. Understanding this distinction can help you interact with AI tools more effectively.
You also explored what makes a prompt effective and how to apply prompt engineering techniques to improve results. By using structured frameworks like CLEAR, PROBE, and PROMPT, you can refine your inputs to guide the AI toward more relevant, accurate, and useful responses. Finally, you examined the limitations of generative AI in retrieving reliable information, such as outdated data, hallucinations, and lack of verifiable sources. These limitations reinforce the importance of thoughtful prompting, careful fact-checking, and choosing the right tools for academic tasks.
Clearly define what you want the AI to do, starting with an action verb (e.g., generate, analyze, summarize). Ambiguous prompts lead to vague or generic responses, while specific instructions produce more relevant results.
AI generates better responses when given the right amount of context. Consider details like the user's needs, goals, and constraints. For example, instead of asking for a workout plan, specify fitness level, goals, and available time per session. Remember to leave out personal identifiers.
Structuring your prompt with goal-focused language helps the AI understand the desired outcome. Instead of saying, Don’t make this too technical, rephrase it as Explain this in simple, beginner-friendly language.
Instead of a broad request, break down multi-step tasks into clear, sequential instructions. For instance, when analyzing user feedback, first ask AI to summarize key takeaways, then categorize them by topic. Telling the AI to ask you questions about the task is also useful. End your prompt with: "Ask me, if you require more information to complete this task".
Even when a response seems well-written, it's important to stay critical. Always check whether the AI's claims are accurate by verifying facts and tracing them back to reliable sources. This is especially important for academic or research-related work, where credibility matters.
These additional resources will help you
Congratulations on completing this module! The quiz is next. Before you begin, make sure you understand the limitations of generative AI for information retrieval, the prompting techniques that can improve response quality, and the strategies you can use to to maintain academic integrity when using AI.
By taking the quiz, you confirm that you've reviewed and completed all the activities in this module.
After the quiz, kindly complete the short survey to help us improve future versions of this module.
The University of Saskatchewan's main campus is situated on Treaty 6 Territory and the Homeland of the Métis.
© University of Saskatchewan
Disclaimer|Privacy