Skip to Main Content

OPAN 6552 – Digital Technologies and Analytics

Professor: Gregory Lyon

What is GAI?

Why use GAI
Good for  Not Good for
Writing models

Domain expertise

Summarizing  Searching propriety databases 
Copy editing Original content  
  Reasoning, like mathematics
 

 Current reality

Warnings

1) Results may contain out-of-date, wrong, and made up information

2) Depending on court cases regarding Copyright, there may be major revisions in the LLM

3) Materials used to train the LLM, do contain bias

4) GAI output is NOT your final project, customize and check everything

Do Use

Critical thinking skills

Generative Artificial Intelligence (GAI) is nothing more than a tool that predicts or creates an output based on the input (prompts) that is used.  It uses Natural Language Processing (NLP) from the Large Language Models (LLM) of text and images it was trained on to develop patterns of output: essays, images, slide content, etc.  The LLM may have a period when their training stopped and have no new information, therefore static.  The prompts are limited in size by how many tokens are used.  Tokens are a breakdown of the words used in the prompts and each GAI has its own limits.  With each output, you can instruct the GAI on how to refine the results.  The best prompts come from domain knowledge and that knowledge should also be used in refining the output.  It is known that some GAI output may contain false facts and incorrect citations, all must checked for accuracy before handing in any assignment.

Some Open GAI tools are below. 

Prompt Engineering

CREATE model developed by David Birss (LinkedIn Learning, 3/15/2023)

Character- what is the role of the GAI?

Request- what you want the GAI to do. Give it context, domain knowledge will be helpful.

Example (Optional)- to set tone or avoid saying what not to do but say what to do instead.

Adjustment- AFTER you submit your prompt, give it more direction on what you want

Type- What Output you want: format, how long, etc.

Extras- “Ask it to show its reasoning” or “Ask me questions before you answer”

Example: 

You are a senior executive in the [industry]. How will the monopoly trial of United States v. Google LLC (2023) in the U.S. District Court for the Eastern District of Virginia Please provide an overview of the court activity in the area of your research, with your analysis of what political and economic forces are at work. Please make a forecast about the likely outcome of the events. Who are the parties of interest? What laws are relevant to affect my industry? Give it to me in a 1000-word memo with a summary and scenarios of possible outcomes. Ask me questions before you answer.

The more descriptive and detailed the prompt is, the better the results.  You are using reinforcement learning from human feedback (RLHF) so use the chat to get what you want. 

Citing Information Generated by ChatGPT/AI

Ask your program about their guidelines for incorporating information generated by ChatGPT and other AI in your papers. When permissible to use that information directly, use the suggestions below.

To cite the informational product generated by ChatGPT or other AI, the recommendation is for the Methodology and/or Introduction of your paper to specify the following:  

  • The prompt you used when utilizing ChatGPT; and 
  • The text that the chatbot produced in response. If the response from ChatGPT is lengthy, please include it in the form of an Appendix. 

As for the in-text citation and the Reference item that will follow this description, apply APA Publication Manual’s templates for citing software (Section 10.10). 

In-text example: 

(OpenAI, 2023) 

Reference example: 

OpenAI. (2023). ChatGPT (May 24 version) [Large language model]. https://chat.openai.com/ 

Creative Commons   This work is licensed under a Creative Commons Attribution NonCommercial 4.0 International License. | Details of our policy