If you have seen phrases like machine learning and deep learning when reading about artificial intelligence (AI), but not really known what they mean, then this webpage is for you! It focuses on the fundamental aspects of AI to help instructors and students recognize how AI tools are built, the different kinds of AI that currently exist, and various ways that AI is used in everyday life, education, and industries. This webpage and the AI 101 webpage, address two AI Literacy competencies (Hibbert et al, 2024; University of Adelaide, 2024), namely, how to:

  1. Design and construct prompts, recognize and apply guidelines and policies for appropriate uses of AI according to context, and document AI use appropriately. (AI 101 - Campus Edition)
  2. Recognize, identify, and distinguish between AI terms, concepts, models, training, software applications, and uses across sectors and disciplines. (AI 102 - AI Fundamentals)

We begin with an overview of what AI is—and is not—followed by examples of AI in industry, and then finish with a curated list of essential AI terms.
 

Artificial Intelligence is…

graphic of AI logo combined with a brain and lines stemming from the AI center

 

Artificial Intelligence (AI) is a term for the overarching eventual goal of building systems that can think and act intelligently like humans (California Miramar University, 2024). Today’s AI systems have not yet achieved that goal, but they can mirror ways that humans think and act in specific circumstances. AI systems use algorithms (i.e., sets of rules used for computer processes), computing models, software platforms, apps, and machinery that enable machines to perform tasks that require human intelligence, such as problem-solving and decision-making. AI systems are used for a range of purposes from completing specific tasks to generating complex, creative content. Most modern AI systems are built using machine learning, which involves programming an AI system to teach itself the rules, patterns, and associations within a data set. These algorithms are programmed with “hyperparameters whose values are set before the machine learning process begins … to affect speed and quality,” (C3.ai, n.d.). That process creates a machine learning model, which derives algorithms for using the data in an AI system.

 

Machine learning models require massive amounts of data that is processed using neural networks within deep learning models to perform multiple iterations of computations on the data. Deep learning models can make use of several algorithms and datasets to complete tasks. They are trained using these four primary categories of algorithms (Coursera, 2023):

  1. Supervised: humans pre-label the data before it goes through machine learning (e.g., video content from cameras on a car).
  2. Semi-supervised: a small amount of data is pre-labelled and the machine uses that to train itself on the remainder of the unlabelled data.
  3. Reinforcement learning: the machine assigns positive and negative values to desired and undesired actions to teach the machine to maximize positive outputs.
  4. Unsupervised: the machine teaches itself without labels or reinforcement.

Deep learning has become progressively more complex over the past decade, incorporating natural language processing (for language content), computer vision (for visual content), and autonomous systems (for robotics) (Coursera, 2024). This has enabled AI systems to become embedded in everyday life and has impacted processes and possibilities across many disciplines and industries, as highlighted in the section of this webpage entitled AI in Disciplines and Industries. 

The next section defines and describes generative AI systems (e.g., ChatGPT) because they are increasingly used by faculty and students for teaching and learning purposes, and are currently being integrated into everyday workplace practices.  

Generative AI is, and is not…

Generative AI (Gen AI) systems are built and trained to “generate new objects that look like the data [they were] trained on” (Zewe, 2023). Transformer technologies are incorporated into machine learning models for Gen AI systems so they can perform natural language processing, which enables the system to output human-like language, behavior, or products. Other common models that Gen AI systems leverage are large language models (for text-based content), Generative Adversarial Networks (GANs) and/or diffusion models (for images and media content). See the AI Terminology section of this webpage for definitions and examples.

 

Generative AI is not an internet search engine like Google or Bing—even when there is an AI summary at the top of the search results. Search engines scour the internet for webpages with information about your query. In contrast, Gen AI systems respond to your query in natural language, generating summaries and explanations that replicates relationships within its training data (some of which could come from internet sites). Gen AI is also known to “hallucinate” (fabricate or make up) a response because it is just replicating patterns in the data, whereas a search engine will not fabricate a website and add it to the list of sources.

 

image of the letters AI floating above a desk with a tablet

 

Generative AI is not a replacement for human expertise. It lacks depth of understanding, contextual awareness, and real-world experience (Willingham, 2019). Some AI proponents suggest that Gen AI can be good at formulaic processes like organizing, summarizing, and comparing to help humans problem-solve, but it can’t anticipate human intentions, adaptability, or emotion. AI cannot fully replicate the nuanced guidance, mentorship, and personalized attention that skilled teachers provide to students because it has no innate agency, goals, values, or ethical reasoning to guide it.

 

Generative AI is not an intelligent critical thinker. In fact, AI can’t “think” at all. Though it can create output from its existing training data, it is not self-directed, cannot intentionally draw conclusions in novel ways (Willingham, 2019), and does not self-reflect on its process or output. “[Humans'] thinking is always tied to our goals, desires and needs... Machines only have the purposes that we give them, and sometimes not even those” (Economist Education, 2024). The Oregon State Ecampus (2023) has a helpful chart that compares AI skills to distinctive human skills, and highlights the critical thinking skills that are uniquely human. It also has examples of how Gen AI could be used to supplement teaching and learning.  

Generative AI is not always a reliable source of information for the following reasons:

  1. Most Gen AI systems are trained on a predetermined set of data within a specific period of time and cutoff date; and they are not equally trained across all domains. While many Gen AI systems do not have access to current events or recent developments in specific fields, some (like Gemini and Bing) do.
  2. Not all Gen AI systems were trained on data that was vetted for quality, credibility, or reliability. Since Gen AI can’t "think," it can’t evaluate the output it creates from that data. It cannot consider context, broader implications, nuances, and background information outside of the prompts its user inputs (Carucci, 2024).
  3. Gen AI systems are widely known to “hallucinate” (fabricate or make up) a response because they put together words or pixels that are statistically closely related to each other. They have no way to verify the accuracy of the output.
  4. Gen AI systems are innately biased because their training data is historical, and they can only perpetuate the patterns and relationships they learn from that dataset. That data and the output generated from it may contain biases and assumptions that humans may not want to perpetuate.  

AI in Disciplines and Industries

Like many other technologies, AI has and will continue to impact most disciplines and sectors of the labor force, sometimes in very controversial ways. Many authors emphasize that the impact of AI on labor will depend on the decisions of policymakers, researchers, and educators (Stropoli, 2023; Georgieva, 2024; Council of Economic Advisors, 2024, Haensch et al., 2023). Instructors are in a position to train students about AI in their discipline and how to leverage AI tools appropriately and effectively for the workforce of tomorrow.

While AI may not entirely replace industries, many jobs will likely be augmented by AI, meaning people will work alongside AI (Dai et al., 2023). Industries and businesses, big and small, benefit from employees who can effectively leverage (or even design) tailored AI technologies to reduce costs while improving processes, practices, and efficiency (TED, 2022; Bombalier, J., n.d.; Bell, E., 2024;  Holdsworth, J., 2024; Quiroz Vazquez & Goodwin, 2024). The examples below provide a few illustrations and details about each can be found in the references section.  
 

Education

K-12 and higher education might be assisted by AI tools designed to track student progress, generate more personalized educational materials or tutoring, and assist multilingual students and students with disabilities. Educators of all disciplines are increasingly feeling the pressures of teaching students to effectively and ethically use AI tools.

Medical Fields

Doctors and other clinical providers use AI tools to aid in clinical decisions, diagnoses, and answering patient questions. In some studies ChatGPT has been shown to perform with relatively high accuracy in clinical decision-making.

Human Resources

AI technologies assist human resources and recruiters in screening applicants and selecting promising candidates based on keywords and other attributes of their resumes.
 

Agriculture

AI enables efficiency in many aspects of the agriculture sector, such as precision farming and crop monitoring through analyzing weather patterns, optimizing water, waste and pesticide usage, and detecting disease.

Finance

The finance industry uses AI to enhance predictive financial models, handle stock trading, explore and manage risk, model personal finance, and streamline customer service.

Security

AI has played a significant role in security for some years (e.g., face recognition), and its role and uses will increase even more for cybersecurity and compliance oversight.

Telecommunications

Many telecommunications providers use AI to monitor and improve their networks and predict problems that might arise in their systems. They also use AI for chatbots, fraud detection, and to automate back-office processes.

AI Terminology

The AI Terminology in this list is a beginner’s guide to the jargon and processes of AI. Read it from top to bottom to build your foundational AI knowledge. You may also want to refer to UC AI Primer online module and the AI 101 webpage for additional terminology.

A term for the overarching eventual goal of building systems that can think and act intelligently like humans. AI systems use algorithms (i.e., sets of rules used for computer processes), hyperparameters (i.e., set values to control things like speed and quality), computing models, software platforms, apps, and machinery that enable machines to perform tasks that require human intelligence, such as problem-solving and decision-making. Today’s AI systems range from narrow purposes, like performing a specific task, to creative purposes like generating audio-visual content.

Systems that are programmed by humans to emulate human behavior by following specific rules and logic within a limited dataset. Symbolic AI can be used within machine leaning. Examples: original versions of Siri, Alexa, bots, and medical diagnosis systems.  
 

A range of approaches to developing algorithms (i.e., computer processes and sets of rules) for statistical data analysis that can subsequently autonomously identify a dataset’s rules and patterns. Humans are involved in the process to build hyperparameters and algorithms, clean data, engineer features, and normalize data types. Machine learning can use Natural Language Processing (for text) and Computer Vision (for media) to identify and classify data. Algorithms used in machine learning require massive amounts of data in the training process.

A subset of machine learning patterned after the interconnections of neurons and synapses in a human brain. After data enters the network’s first layer, it goes through a hidden layer of nodes where calculations adjust the strength of connections before sending the data to an output layer.

A transformer is a neural network architecture that uses an encoder to find relationships in sequential data (like text), and a decoder to produce outputs that follow those relationships.
 

Neural networks with multiple hidden layers that enable computers to autonomously discover patterns and make decisions from vast amounts of unstructured data. Be critical: The decisions outputted by deep learning models are often very difficult to interpret as there are so many hidden layers doing different calculations that are not easily translatable into recognizable rules.

Deep learning models that are pretrained on extensive text-based datasets and use a transformer to learn patterns in the data so it can output human-like text. (Guardian) Examples: Chat-based generative pre-trained transformer (GPT) models like ChatGPT and Gemini.

Deep learning architecture usually trained on audio-visual data that uses two parallel neural networks to generate content similar to its training data:

(1) Generator: modifies an input data sample to create new data.

(2) Discriminator: Determines if the generated data is real or fake. The two networks compete against each other to improve the generated data until the discriminator can no longer tell the difference between the fake and original data. Examples: generating video game characters, deep fakes, and medical imaging. (More information on Generative Adversarial Networks (GAN).) 

Deep learning model usually trained on audio-visual data in which the neural network is trained to add random “noise” to the data. The model learns to reverse the process of adding noise to the data, gradually transforming the noise into a structured output. Example: Dall-E and Stable Diffusion. (More information on diffusion models.)
 

AI systems that are built using deep learning models and trained to generate new objects that look like the data it was trained on. Depending on the combination of the system’s data and deep learning models, it can generate content such as text, images, music, videos, or 3D models. Be critical: As a GenAI receives input in the form of prompts and feedback from users, it adds that data to its training model and continues learning from it. Users can prevent their data from being added to the training data in their privacy settings.

References

Bell, E. (2024). How AI Is Used in Business: AI is having a significant impact on the business world, affecting jobs, workers, firms, and industries. Investopedia.
https://www.investopedia.com/how-ai-is-used-in-business-8611256

Bombalier, J. (n.d.). The Competitive Advantage of Using AI in Business. Florida International University Chapman Graduate School.
https://business.fiu.edu/academics/graduate/insights/posts/competitive-advantage-of-using-ai-in-business.html

C3.ai. (n.d.). Glossary: Data Science - Hyperparameters. C3 AI. Retrieved December 17, 2024, from
https://c3.ai/glossary/data-science/hyperparameters/

California Miramar University. (2024, May). Artificial Intelligence vs. Machine Learning | CMU [Education]. Artificial Intelligence vs. Machine Learning: Differences, Uses & More.
https://www.calmu.edu/news/artificial-intelligence-vs-machine-learning

Carucci, R. (2024, February 6). In The Age Of AI, Critical Thinking Is More Needed Than Ever. Forbes.
https://www.forbes.com/sites/roncarucci/2024/02/06/in-the-age-of-ai-critical-thinking-is-more-needed-than-ever/

Cleveland Clinic (2024). How AI Is Being Used to Benefit Your Healthcare. Cleveland Clinic HealthEssentials News and Innovation.
https://health.clevelandclinic.org/ai-in-healthcare

Coursera. (2023, November 29). Machine Learning Models: What They Are and How to Build Them. Coursera.
https://www.coursera.org/articles/machine-learning-models

Coursera. (2024, March 26). 4 Types of AI: Getting to Know Artificial Intelligence. Coursera.
https://www.coursera.org/articles/types-of-ai

Council of Economic Advisors (2024). Potential Labor Market Impacts of Artificial Intelligence: An Empirical Analysis.

Dai, Y. et al. (2023). Reconceptualizing ChatGPT and generative AI as a student-driven innovation in higher education. 33rd CIRP Design Conference Proceedings.

Economist Education. (2024, July 16). AI can’t think critically, but people must. Economist Education Insights | Interviews.
https://education.economist.com/insights/interviews/ai-cant-think-critically-but-people-must

Fino, M. and Downie, A. (2023). What is artificial intelligence (AI) in finance? IBM Think.
https://www.ibm.com/topics/artificial-intelligence-finance

Georgieva, K (2024, January 14). AI Will Transform the Global Economy. Let’s Make Sure It Benefits Humanity. IMF Blogs
https://www.imf.org/en/Blogs/Articles/2024/01/14/ai-will-transform-the-global-economy-lets-make-sure-it-benefits-humanity

Haensch, A. C., Ball, S., Herklotz, M., & Kreuter, F. (2023, October). Seeing ChatGPT through students’ eyes: An analysis of TikTok data. In 2023 Big Data Meets Survey Science (BigSurv) (pp. 1-8). IEEE.

Hibbert, M., Altman, E., Shippen, T., & Wright, M. (2024, June 3). A Framework for AI Literacy [Education]. EDUCAUSE Review.
https://er.educause.edu/articles/2024/6/a-framework-for-ai-literacy 

Holdsworth, J. (2024). The most valuable AI use cases for business. IBM Think.
https://www.ibm.com/think/topics/artificial-intelligence-business-use-cases

James Madison University Libraries (2023). Artificial Intelligence (AI) in Education. Research Guides.
https://guides.lib.jmu.edu/AI-in-education

Johar, V. (2022). Artificial Intelligence In Hiring: A Tool For Recruiters. Forbes.
https://www.forbes.com/councils/forbesbusinesscouncil/2022/06/10/artificial-intelligence-in-hiring-a-tool-for-recruiters/

Kropov, V. (2024, October 23). Transformative applications of AI in telecommunications. AI in Telecommunications: Uses, Challenges, Best Practices.
https://www.n-ix.com/ai-in-telecommunications/

Los Angeles Pacific University (2023). Revolutionizing Healthcare: How is AI being Used in the Healthcare Industry? Los Angeles Pacific University Career Blog.
https://www.lapu.edu/ai-health-care-industry/

Massachusetts General Brigham (2023). ChatGPT Shows ‘Impressive’ Accuracy in Clinical Decision Making. Massachusetts General Brigham Newsroom.
https://www.massgeneralbrigham.org/en/about/newsroom/press-releases/chatgpt-shows-impressive-accuracy-in-clinical-decision-making

Oregon State Ecampus. (2023). Bloom’s Taxonomy Revisited. Oregon State Ecampus.
https://ecampus.oregonstate.edu/faculty/artificial-intelligence-tools/blooms-taxonomy-revisited.pdf

Restrepo, D., & Abdulnour, R. (2023). How Is AI Used in Health Care? Massachusetts General Brigham Newsroom.
https://www.massgeneralbrigham.org/en/about/newsroom/articles/how-is-ai-used-in-health-care

Rao, A., Pang, M., Kim, J., Kamineni, M., Lie, W., Prasad, A. K., ... & Succi, M. D. (2023). Assessing the utility of ChatGPT throughout the entire clinical workflow: development and usability study. Journal of Medical Internet Research, 25, e48659.

Ross, E. (2023). Embracing Artificial Intelligence in the Classroom: Generative AI tools can reflect our failure of imagination and that is when the real learning starts. Harvard Graduate School of Education Usable Knowledge.
https://www.gse.harvard.edu/ideas/usable-knowledge/23/07/embracing-artificial-intelligence-classroom

Ruiz, P., & Fusco, J. (2024, March 31). Glossary of Artificial Intelligence Terms for Educators – CIRCLS [Organization]. Glossary of Artificial Intelligence Terms for Educators.
https://circls.org/educatorcircls/ai-glossary

Russell, R. G., Lovett Novak, L., Patel, M., Garvey, K. V., Craig, K. J. T., Jackson, G. P., Moore, D., & Miller, B. M. (2023). Competencies for the Use of Artificial Intelligence–Based Tools by Health Care Professionals. Academic Medicine, 98(3), 348.
https://doi.org/10.1097/ACM.0000000000004963

Stropoli, R (2023, November 14). A.I. Is Going to Disrupt the Labor Market. It Doesn’t Have to Destroy It.: The decisions of companies, governments, and educators will help to shape the ultimate outcomes of the A.I. revolution.” Chicago Booth Review
https://www.chicagobooth.edu/review/ai-is-going-disrupt-labor-market-it-doesnt-have-destroy-it

TED (2022, October 13). How AI Could Empower Any Business | Andrew Ng [video]. YouTube.
https://www.youtube.com/watch?v=reUZRyXxUs4

TED (2023, May 1). How AI Could Save (Not Destroy) Education | Sal Khan [video]. YouTube.
https://www.youtube.com/watch?v=hJP5GqnTrNo&ab_channel=TED

UCTV (2024, July 24) Al for Security, Security for Al with Christopher Kruegel.
https://www.uctv.tv/shows/Al-for-Security-Security-for-Al-40086

University of Massachusetts Amherst (2024). AI in Finance & Banking: 11 Ways It’s Changing the Industry. University of Massachusetts Amherst Isenberg School of Management Blog.
https://iconnect.isenberg.umass.edu/blog/2024/10/25/ai-in-finance-banking-11-ways-its-changing-the-industry/

University of Adelaide, L. (2024, August). Artificial Intelligence Literacy Framework. University of Adelaide.
https://www.adelaide.edu.au/library/system/files/media/documents/2024-10/ua31352_ai-literacy-framework-doc-v2.pdf

Willingham, D. T. (2019). How to Teach Critical Thinking. New South Wales Department of Education.
https://education.nsw.gov.au/teaching-and-learning/education-for-a-changing-world/thinking-skills/how-to-teach-critical-thinking

Willmore, J. (2023). AI education and AI in education. NSF Science Matters.
https://new.nsf.gov/science-matters/ai-education-ai-education

Windley, D (2021). Is AI The Answer To Recruiting Effectiveness? Forbes.
https://www.forbes.com/councils/forbeshumanresourcescouncil/2021/06/16/is-ai-the-answer-to-recruiting-effectiveness/

William and Mary University (2024). The Future of Finance: AI, Machine Learning and Predictive Analytics. William and Mary Raymond A. Mason School of Business Online Business Blog.
https://online.mason.wm.edu/blog/the-future-of-finance-ai-machine-learning-predictive-analytics

Williamson, B. (2024). The Social life of AI in Education. International Journal of Artificial Intelligence in Education, 34(1), 97–104.
https://doi.org/10.1007/s40593-023-00342-5

Zewe, A. (2023, November 9). Explained: Generative AI [Education]. MIT News | Massachusetts Institute of Technology.
https://news.mit.edu/2023/explained-generative-ai-1109