AI and Information Literacy: Prompt Effectively

Learn about the intersection of Artificial Intelligence (AI) and information literacy.

Introduction: Prompt Effectively

What is Prompt Engineering?

There exist many descriptions of prompt engineering, but a very succinct one is provided by Dany Liu from the University of Sydney in an article. Liu describes prompt engineering as: "When you use tools like ChatGPT, you provide it a ‘prompt‘, which the software sends to a ‘model‘ (the actual AI), and the model then produces a ‘completion‘ which is shown to you through the tool." Imagine prompt engineering as something akin to designing a website, where you have the ability to choose from several coding languages like HTML, CSS, and Bootstrap to achieve the ideal design according to your target audience and communication goals.

Google Search versus Prompt

Google	Prompt
Information searching	Language generation
Keyword based	Context based
Limited interaction	Conversational

Why it's important to learn

A growing number of apps require prompting skills.

Prompting helps you learn, finish new tasks, and be creative

Skilled prompters prevent AI app hallucinations, erroneous information, and other ethical issues.

Before you begin

Protect your personal information and think carefully before using a machine learning artificial intelligence tool. For example, consider a section of Google's privacy policy. Did you realize that Google has human reviewers? Note what they've bolded:

How human reviewers improve Google AI

To help with quality and improve our products (such as generative machine-learning models that power Gemini Apps), human reviewers read, annotate, and process your Gemini Apps conversations... Please don't enter confidential information in your conversations or any data you wouldn't want a reviewer to see or Google to use to improve our products, services, and machine-learning technologies.

Source: “Gemini Apps Privacy Hub - Gemini Apps Help.” Accessed July 22, 2024. https://support.google.com/gemini/answer/13594961?visit_id=638572722688932795-429543449&p=privacy_help&rd=1#privacy_notice.

Things you can ask GenAI to do

Proofread a document

Debug or generate code

Check the tone of something you wrote

Find information on a topic you know a lot about

Find information on a topic you know very little about

Compose a Tweet based on a longer paragraph of text

Summarize notes from a meeting

Simplify a concept (e.g. "explain [concept] to me as if I'm 12 years old")

Make a plan (e.g. "recommend a list of roadside attractions where I should stop when driving from College Park, Maryland to New York City")

Ask for coaching or have a longer conversation (e.g. "Act as an interviewer for [job]. Please ask me one question at a time and wait for my response each time. Have the questions start easy and get progressively more technical.")

Tips for writing prompts

Principles for good prompting

Craft precise and unambiguous instructions.
Use the conversational interface on many generative AI tools to refine its outputs.

Specificity is key! The model will make an assumption on any point you don’t clearly explain.

Try and then refine

There is no hard science to good prompting. Instead, approach it as a process of trial and error:

Formulate a relevant prompt.
Note where the completions/responses fall short of expectations.
Enhance the clarity of the prompt. You could refine your prompts using a set of exemplars, such as the ones found on this site.
Repeat the process to iteratively improve your prompt's effectiveness.

Remember that as you chat with generative AI, the model will learn more about what you need. Use the conversational turns to help the AI improve its outputs.

Also remember that simple prompts lead to simple (and similar) completions.

Once you've got the hang of basic prompting, you might like to look at more advanced techniques to get the most out of generative AI models.

This section was adapted from a course created by the University of Sydney. This course content is offered under a CC Attribution Non-Commercial license.

Frameworks

Frameworks allow you to organize your prompts more efficiently. To improve your ability to use the tools and obtain more accurate and better answers, follow well-organized and clear prompts. They can assist you in developing a precise question. You will find that tools like ChatGPT and Microsoft Copilot come in handy more frequently than you think, and you will be surprised at all of the things they can do.

Framework	Meaning	Explanation	Example
TAG	Task, Action, Goal	Define the task, define the job to be done, and explain end goal.	Task: summarize the main points of this journal article on depression and obesity. Action: condense the answer into a precise paragraph. Goal: Provide a clear understandable explanation of the key arguments of the paper.
CARE	Context, Action, Result, Example	Give background information , define the action needed, define your goal, give some examples	Context: You are a helpful assistant who writes Python code. Action: Write a function to calculate the factorial of a given number. Result: The function should return the factorial as an integer. Examples: Input:5, Output: 120
RACE	Role, Action, Context, Explanation	Specify role, mention action needed, provide background information, describe your outcome	Role: You are a pediatric dentist Audience: pediatric patients between 5 -10 years old. Context: Write a one page summary about the importance of oral hygiene. Example: Focus on standard from the American Dental Association

Additional resources

Cited Work

Prompt engineering for educators – making generative AI work for you – Teaching@Sydney. (2023, April 27). https://educational-innovation.sydney.edu.au/teaching@sydney/prompt-engineering-for-educators-making-generative-ai-work-for-you/