Defining AI
Through machine learning, AI is a computer "program" that imitates human like interactions. By training an AI with thousands of books and articles, most AI has encyclopedic knowledge and can answer basic questions. Most AI is also trained with large data dumps of social media to help the AI with creating something akin to natural or current language. With a series of prompts most AI is capable of creating long form answers to questions and prompts. Those long form answers may be paragraphs or entire essays.
Because of the training of AI it is prone to certain problems.
Most of this is called "hallucinations", but it is the AI relying on its training. "Garbage in. Garbage out."
Additionally, the AI is programmed to put words together based on probability of what should come next. It has learned what typically follows certain words or phrases. So, AI tends to be generic lifeless writing because it writes what is expected. It isn't creative with expressions.
Some Assignment Ideas to Stop the Bot (OK, really just speedbumps for the students)
List of Current AI Checkers (8/23)
The University also uses TurnItIn. TurnItIn has a built in originality checker and an AI detector. As an instructor you need to enable those items in Canvas
You may detect it yourself.
Look for the following:
To watch for (possible false positives):
For More About AI
Much of this comes from: https://sites.google.com/view/practical-information-literacy/beating-ai - created by Steve Jung
On that site are links to two PowerPoints on AI.
The top PowerPoint was delivered to the faculty of HIU for the Spring faculty in-service meeting in March 2023.
The other PowerPoint was delivered as a sponsored webinar for Atla (formerly the American Theological Library Association) and emphasizes the changes need to academic integrity statements because of artificial intelligence.
For a Glossary of Terms