Glossary of Terms
Here are some different Artificial Intelligence Glossaries that can help explain complex or overlapping terms.
- NIST The Language of Trustworthy AI: An In-Depth Glossary of TermsThe National Institute of Standards and Technology, from the U.S. Department of Commerce, March 22, 2023.
- The New York Times Artificial Intelligence GlossaryAdam Pasick, The New York Times, March 27, 2023.
About this Guide
The use of Artificial Intelligence in scholarship is an ever evolving topic, with many nuances and subjective stances. As of now, College of the Holy Cross does not have an official policy on the use of AI in course work. This means that you are responsible for checking your class syllabi and speaking with your professors about their policy on AI use in their own courses. This guide will help you understand what Artificial Intelligence really is, how it works, and show you how to navigate the different applications of AI to your course work.
Your Ideas
This guide is a work in progress! If you have a resource, a topic, or an AI-related update that you think should be represented here, let us know:
What is Artificial Intelligence?
What is Artificial Intelligence?
Artificial Intelligence or AI was first defined by Stanford Processor John McCarthy in 1955 as “the science and engineering of making intelligent machines." Today, we know AI as the science and engineering of machines that can perform tasks requiring human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages. While these machines are taught to mimic human intelligence, none of these machines are sentient or self aware; they are being programed with an established skill set and the ability to adapt when given knew information or instructions.
Types of Artificial Intelligence
AI exists in many different formats and implementations, which can make the classification systems for AI confusing and difficult to determine. The broadest classification system for AI is functional AI, AI tools that have been developed and are actually usable, versus theorized AI which includes AI models that scientists believe could be developed but so far have been unsuccessful. This guide will focus on actual models of AI available for use.
Functional AI can be sorted into two categories based on their capabilities:
Reactive AI is programmed to provide a predictable output based on the input it receives. Reactive machines always respond to identical situations in the exact same way every time, and they are not able to learn actions or conceive of past or future. Some examples of reactive AI include the spam filter on your email and the Netflix recommendation engine.
Limited Memory AI is programmed with a limited memory so that it can learn from past data to provide better results. This type of AI is able to provide results derived from both an existing general knowledge base and observational data or inputed data from the user. This type of AI is not able to add inputed data to it's existing knowledge base- meaning it does not learn over time or retain any information it did not previously have stored in it's programming. This is the most widely used type of AI, especially in Academia.
In recent years, higher education has become particularly concerned with generative AI, a subset of limited memory AI. Generative AI is a type of artificial intelligence which is capable of producing content -- code, text, images, etc. -- by combining user input with it's existing knowledge base, called a large language model. Large Language models are a collection of large data sets that supply AI with it's knowledge base and train AI to generate content. Some examples of generative AI tools or sites that use generative AI are:
- Chat GPT
- Grammarly
- Socratic
- Quizlet
- Duo Lingo