Skip to Main Content

Beginner's Guide to Generative AI

USC AI Policy

As of August 2025, USC does not have an official policy on using generative AI for assignments. Currently, the library recommends that students learn about AI and the various citation methods. Please check with your professor or department regarding their stance on using AI as a tool/resource.

Defining Generative AI

Generative Artificial Intelligence (GenAI) is defined by Khan (2025) as:

... a type of Artificial Intelligence that focuses on creating new content or data. It can generate realistic images, videos, texts, and sounds that may appear to be created by humans. Unlike traditional AI, which analyzes and interprets data, Generative AI goes a step further by producing new, original output.

Examples of what GenAI can produce, but are not limited to:

  • Writing and Text Generation
  • Deepfakes
  • Music Composition
  • Art Creation
  • Product Design

Khan, Richard R. AI Glossary : Demystifying 101 Essential Artificial Intelligence Terms for Everyone. First edition. CRC Press, Taylor & Francis Group, 2025. Taylor & Francis Ebooks, https://www.taylorfrancis.com/books/mono/10.1201/9781003600268/ai-glossary-richard-khan.

AI Terms/Jargons

  • Algorithm

    A set of rules or processes used by an AI system [on the backend] to conduct tasks for the user.5

  • Artificial Intelligence

    Intelligence displayed or simulated by technological means. Often it is assumed that “intelligence” in this definition means: considered intelligent by the standard of human intelligence, the sort of intelligent capacities and behavior that humans display. The term can refer to the science or to the technologies, for example, learning algorithms.2

  • Bias

    Discrimination against or in favor of particular individuals or groups. In the context of ethics and politics, the question arises whether a particular bias is unjust or unfair.2

  • Chatbot

    Powered by text generation, chatbots can provide customer support, answer questions, or simulate human-like interactions in various applications.4

  • Copilots

    A tool or system designed to assist software developers by suggesting code, correcting errors, or providing recommendations based on the context of the code.4

  • Data Mining

    The term is misleading because the goal is the extraction of patterns from the data - the analysis of data - not the extraction of data itself.2 

  • Ethical [Trustworthy] AI

    AI that can be trusted by humans. Conditions for such trust can refer to (other) ethical principles such as human dignity, respect for human rights, and so on, and/or to social and technical factors that influence whether people will want to use the technology. The use of the term “trust” with regard to technologies is controversial.2

  • Generative AI

    Aka GenAI -- A type of AI that creates new content or data. It learns from existing data, understands the patterns and styles, to generate realistic images, videos, text, and sounds that appear to be made by a human.4

  • Hallucination

    Misinformation or made-up information based on a pattern that the AI model has learned as part of its training. For example, the model could create references that do not actually exist.3

Note: If a user asks a GenAI bot to do something it is not designed to (ex: asking ChatGPT to find academic articles online), this is considered user error, not a hallucination.
  • Large Learning Model (LLM)

    A type of AI program that has been trained on enormous amounts of text data. This training allows the model to understand, predict, and generate human language with a high degree of proficiency.4

  • Machine Learning

    A machine or software that can automatically learn: not in the way humans learn, but based on a computational and statistical process. Feeding on data, learning algorithms can detect patterns or rules in the data and make predictions for future data.2

  • Neural Network

    A method in artificial intelligence (AI) that teaches computers to process data in a way that is inspired by the human brain.6

  • Natural Language Processing (NLP)

    The capacity of a computer to understand human speech and writing used in various language-related tasks.1

  • Pattern Recognition

    The identification of patterns or rules in data - algorithms then use those patterns or rules to explain the data and make predictions for future data.2

  • Prompt

    The initial input text or instructions given to a model to generate new content based on that starting point. It provides context and guides the model's output. The prompt can be a few words or sentences that set the tone or specify the desired content.3

  • Prompt Engineering

    When a user creates a prompt with specific words, phrases, or symbols to clearly insruct the AI model towards a desired output.7

  • Sustainable AI

    AI that enables and contributes to a sustainable way of living for humanity and does not destroy the ecosystems on Earth on which humans (and also many nonhumans) depend.2

References

  1. Boddington, Paula. AI Ethics : A Textbook. Springer, 2023.

  2. Coeckelbergh, Mark. AI Ethics, MIT Press, 2020. ProQuest Ebook Central, http://ebookcentral.proquest.com/lib/southcarolina/detail.action?docID=6142275.

  3. “Key AI Terms Glossary.” Walden University Academics - Artificial Intelligence, Updated April 13, 2025. https://academics.waldenu.edu/artificial-intelligence/glossary.

  4. Khan, Richard R. AI Glossary : Demystifying 101 Essential Artificial Intelligence Terms for Everyone. First edition. CRC Press, Taylor & Francis Group, 2025. Taylor & Francis Ebooks, https://www.taylorfrancis.com/books/mono/10.1201/9781003600268/ai-glossary-richard-khan.

  5. “What Is a Machine Learning Algorithm?” IBM, Updated April 17, 2025. https://www.ibm.com/think/topics/machine-learning-algorithms.

  6. “What Is a Neural Network?” Amazon Web Services, Updated July 3, 2025. https://aws.amazon.com/what-is/neural-network/.

  7. “What Is Prompt Engineering?” Amazon Web Services, Updated July 3, 2025. https://aws.amazon.com/what-is/prompt-engineering/.