AI and Academic Integrity

Warning

A picture of a "pointer" dog in front of a blackboard indicating that users of AI should ensure that they have talked with their instructors to ensure that AI is allowed in the class.

What is Generative Artificial Intelligence?

Generative artificial intelligence (AI) is a category of web-based tools that use algorithms, data, and statistical models to draw reasonable inferences to create content of its own (e,g., text, images, etc.). They are not search engines but rather trained chatbots. Using a prompt, a chatbot strives to fill in the next missing content piece, "what one might expect" (Wolfram). 

These tools use large language models to provide bots with the data they need to reply to a prompt you have given it appropriately. For example, when ChatGPT writes a response to a prompt, it provides text based on what words came before and what is the most likely next word. Because AI uses natural language and computes so quickly, it can often seem like the chatbot is, in fact, intelligent. 

The field of AI is changing at a rapid pace. We know that these generative tools help users synthesize information and create content (code, essays, art, music, etc.). However, these tools can also "hallucinate", or make up facts or sources and create biased content. 

Be sure to make sure it is ethical to use AI  and fact-check any content and sources you plan to use in the work you share with others or publish that has been generated by AI.

Generative AI in Action

This image was generated using Craiyon and the prompt "an image of an arkansas razorback.

It is important to use critical thinking to evaluate and interpret AI-generated results. This issues are easy to see in a graphic instance - although similar issues do occur with text AI as well.

Looking at these AI-generated images, and comparing them to what the University of Arkansas mascot actually looks like.  While some of the images may somewhat resemble Tusk, some have blurred words and resemble cows, dogs or wolves

While we can't determine precisely how the model went awry, it is likely that the underlying training data  included general information on what is a razorback, that the razorback is a mascot, and that the colors of the University of Arkansas included red, the model seemed to construct an image that included elements of other college mascots or animals of similar shapes.

 

An image from Craiyon of the razorback hog with many variations that are not he true hog