Subject Guides
Generative Artificial Intelligence
An introduction to generative AI, guidance on using AI tools, and additional resources
Guide Contents
Prompting AI
There are many frameworks for writing effective prompts when using generative AI tools. 'Effective' prompts mean more than just trying to get relevant output from the gen AI tool. Effective prompt writing is also important to: reduce or acknowledge bias, acknowledge limitations of the tool and control the type of interaction you have with the tool.
The tabs in this section introduce different frameworks that can be used to write effective prompts for AI. But, beyond writing prompts there are other things to be aware of when using gen AI tools.
- Ensure the tool matches the purpose. For example, don't rely on Chat GPT to create accurate citations and information although there are tools that can generate accurate citations and summaries of actual articles
- Understand the limitations, rules and rights you have, or don't have, when using any AI tools
- Create effective prompts
- Evaluate the output carefully not only checking for accuracy, but what the output reveals about your original prompt.
- When you ask a Gen AI tool to write an outline from a paragraph it will create the outline, but it won't tell you if your paragraph and initial thesis is too broad, weak, or relevant. However, the outline may show that if the outline seems broad and unspecific, does not seem relevant to the topic or has a very weak argument/structure
- Iterate. Prompt and evaluate, prompt and evaluate and prompt again. Gen AI tools are chatbots and often need additional prompts, examples or information in order to create a refined response.
The CLEAR framework was created by Leo S. Lo specifically to address the need of creating frameworks to introduce students to effectively working with and understanding the limitations of AI tools.
The CLEAR framework:
- Concise: brevity and clarity in prompts
- Logical: structured and coherent prompts
- Explicit: clear output specifications
- Adaptive: flexibility and customization in prompts
- Reflective: continuous evaluation and improvement of prompts
The PROMPT framework was created by Sarah Hartman-Caverly a Librarian Penn State Berks. The framework focuses more clearly on providing role, or persona, to the chat tool. Persona's are ideal to act as editors, teachers or reviewers on top of the 'friendly chatbot' persona already built into tools like Chat GPT.
- Persona - assign a role
- Requirements - define parameters for output
- Organization - describe the structure of the output
- Medium - describe the format of the output
- Purpose - identify the rhetorical purpose and intended audience
- Tone - specify the tone of output (ex: academic)
