Skip to Main Content

Artifical Intelligence

Introduction to AI

Commonly Use Terms

  • Artificial Intelligence (AI) - Technology focused on enabling computers to make decisions, problem-solve, or discover patterns.
  • Algorithm - The rules an artificial intelligence system uses to make decisions.
  • Machine Learning (ML) - A field of study focused on developing algorithms.
  • Neural networks - A subset of machine learning algorithms where interconnected nodes process and analyze data.
  • Deep Learning - A subset of neural networks with the ability to recognize more complex patterns.
  • Large Language Model (LLM) - A type of foundational model trained with massive data sets.
  • Generative AI (GenAI) - A type of machine learning that generates content.

Ruiz, P., & Fusco, J. (2024, March 31). Glossary of Artificial Intelligence Terms for Educators. Glossary of Artificial Intelligence Terms for Educators – CIRCLS 

The image below visualizes how the different components of technology work together to build a final product. 

An image titled "Defining Generative AI" with the subtitle "to understand generative artificial intelligence (GenAI), we first need to understand how the technology builds from each of the AI subcategories listed below." Below, there is one big circle with three progressively smaller circles inside. From biggest to smallest it is Artificial Intelligence, Machine Learning, Deep Learning, then Generative AI.

AI for Education. (2023). Defining Generative AI. Generative AI Explainer — AI for Education

Further Reading

IBM. (2024, August 9). What is artificial intelligence? What Is Artificial Intelligence (AI)? | IBM 

Types of AI

At a broad level, AI can be divided into three categories based on capabilities: 

  1. Narrow AI – Trained to perform a single or narrow task, may also be referred to as Weak AI (i.e., Siri or ChatGPT). 
  2. General AI – Capable of completing a variety of tasks based on previous learning without the need for training, may also be referred to as Strong AI. 
  3. Super AI – Capable of intelligence surpassing human capabilities, may also be referred to as artificial superintelligence. 

Currently, only Narrow AI exists. Both General and Super AI are theoretical concepts.  

Additionally, AI can be divided into four categories based on functionality: 

  1. Reactive Machine AI – Systems with no memory, designed to perform a specific task based on presently available data.
  2. Limited Memory AI - Systems with recall to decide on a course of action most likely to help achieve a desired outcome.
  3. Theory of Mind AI - A theoretical system able to understand the thoughts and emotions of other entities.
  4. Self-Aware AI - A theoretical system able to understand its own internal conditions and traits along with human emotions and thoughts.

IBM. (2023, October 23). Understanding the different types of artificial intelligence. https://www.ibm.com/think/topics/artificial-intelligence-types 

For more information, view IBM's The 7 Types of AI - And Why We Talk (Mostly) About 3 of Them. (CC and transcript available on YouTube)

Applications of AI

Human-AI Activity

Description 

Examples
Content creation generating new artifacts such as video, narrative, software code, synthetic data. subtitle creation; text-to-image

Content synthesis

 combining and/or summarizing parts, elements, or concepts into a coherent whole. converting doctors’ unstructured notes; summarizing a book
Decision making selecting a course of action from among possible alternatives in order to arrive at a solution. buy/sell financial decision
Detection identifying, by careful search, examination, or probing, the existence or presence of [something]. detect cybersecurity threats
Digital assistance acting as a personal agent for understanding and responding to commands and questions, and carrying out requested tasks in a conversational manner. reminders from smart assistants (e.g., Siri, Amazon Echo, Google Assistant, Bixby)
Discovery finding, recognizing, or unearthing something for the first time. drug discovery and production
Image analysis recognizing attributes within digital images to extract meaningful information. medical diagnostics
Information retrieval/search finding information about specific topics of interest. speed the search for stable proteins used in drug development, biofuels, and food production
Monitoring observing, checking, and watching over the process, quality, or state of [something] over time to gain insights into how [something] is behaving or performing. wildfire monitoring

Performance improvement

 improving quality and efficiency of the intended outcomes. graph analytics; increasing efficiency and scalability for graph computing
Personalization designing and tailoring [something] to meet an individual's characteristics, preferences, or behaviors. sales content personalization and analytics
Prediction forecasting the likelihood of a future outcome. sales forecasting; weather forecasting
Process automation performing repetitive tasks, removing bottlenecks, reducing errors and loss of data, and increasing efficiency of a process. automating administrative tasks
Recommendation suggesting or proposing a manageable set of viable options to aid decision-making. customer service response suggestions; purchase recommendations; content recommendations
Robotic automation* using physical machines to automate, improve, and/or optimize a variety of tasks. intelligent robots in surgery
Vehicular automation* automating physical transportation of goods, instrumentation and/or people. self-driving cars/trucks/trains; drones; spacecraft; airplanes

*Robotic automation and vehicular automation involve physical embodiment and represent a different level of abstraction that the other activities in the taxonomy but are included for completeness.

Theofanos, M., Choong, Y., & Jensen, T. (2024, March). AI use taxonomy: A human-centered approach. NIST. doi.org/10.6028/NIST.AI.200-1