Artificial Intelligence in Teaching & Learning

Some Ideas for How GenAI Can Help with School

There are many options for using A.I. in education to help the learning process, including:

• Ask it to explain difficult concepts in a simple way
• Use it to generate ideas for a paper topic
• Get feedback on grammar and spelling
• Ask it to explain why you got a question right or wrong

The University of Sydney, in Australia, offers an AI in Education resource for students, built by students, with ways you can use generative artificial intelligence productively and responsibly in school.

They also point out that GenAI can help you create a resume and prepare for a job interview. 

CLEAR Framework

When using generative AI, a prompt is your question, request, or other input. The CLEAR Framework by librarian and professor Leo S. Lo provides guidance about how to approach AI prompting. 

This guide from Georgetown University gives examples of putting the CLEAR Framework into action for research. 

Elements of a Prompt

It's possible to simply type a question into a tool like ChatGPT, but you probably won't get a very helpful response. That's because the AI needs you to provide specific information and context to address the information you're seeking or the task you hope to complete.

To increase the chances of getting a helpful response, include the following elements in a prompt for generative AI:

• Instruction 
  • The specific task or instruction you want the AI tool to perform
  • Examples: List, summarize, explain...
• Input Data 
  • The input or question you want a response for
  • Examples:
    • A search strategy for an undergraduate nursing research paper
    • Three potential topics for an English paper on The Yellow Wallpaper
• Context 
  • Information that can steer the model to better responses
  • Examples:
    • Write for a college student reading level
    • Write for a middle school reading level
• Output Indicator 
  • The type or format of the output
  • Examples:
    • A short paragraph
    • A list in ascending difficulty

Read more about prompt engineering in the Prompt Engineering Guide written by technology consultants from an organization called DAIR.AI (Democratizing Artificial Intelligence Research, Education, and Technologies).

Note: Prompt engineering can get very technical, but it doesn't have to! Don't be intimidated. Try it, experiment, see what works and what doesn't, and keep on learning!  

Sample Prompts

To practice with generative AI, try these prompts as-is, or substitute your own topics.

• I'm a college student studying botany. Can you help me understand photosynthesis?
• Write a multiple-choice quiz for college students about how the circulatory system works. Create 12 questions with 4 possible answers each. Include an answer key.
• I'm preparing for a job interview for a human resources position at the University of Hawaii. Could you act as the interviewer and ask me some questions?
• I'm a college student and I need to find research articles for my nursing class. The topic is: what screening should be done for perinatal maternal mental health? 
• I need to find research articles in my library's databases on the topic of gun control. Can you suggest other search terms that would get me good results?

Be Smart: GenAI Has Downsides, Too

There are several limitations and areas of concern with ChatGPT and other GenAI, such as:

• Inconsistency: Responses (output) will change if you alter the prompt even slightly, or if the same prompt is used multiple times.
• Incorrect information: GenAI sometimes presents plausible-sounding but incorrect information, since verifying the truth of its sources is not currently part of training. Even when it provides links to sources, it's unable to avoid basic lapses in logic due to how it is built.
• Biased information is also a concern. The prejudices of society are reflected in the data used to train AI systems, even when this is not intentional. This means that tools such ChatGPT "cannot recontextualize or independently seek out new information that contradicts their built-in assumptions” (Williams, 2023). 

Be skeptical and use these tools with care!