Topic
As Gonzaga University expands its use of Artificial Intelligence (AI) tools, it is important for students, faculty, and staff to understand key AI terminology. This reference guide defines common terms to support informed and responsible use of AI across campus.
Terminology
Artificial Intelligence (AI)
Artificial Intelligence (AI) is a computer system that can imitate human capabilities such as comprehending language, making decisions, translating between languages, and learning from experience. AI systems work by running vast collections of data through algorithms — sets of instructions — to create models that can automate tasks typically requiring human intelligence. Sometimes people specifically engage with an AI system directly, such as asking a chatbot for help, but AI also operates in the background in everyday tools, such as suggesting words while typing or recommending content based on preferences.
Generative AI
Generative AI leverages the power of large language models (LLMs) to create new content — such as text, images, music, videos, and code — rather than simply retrieving or summarizing existing information. It learns patterns and structures from training data and generates something similar but new. Generative AI can be used to create art, write stories, design products, and assist with administrative tasks. It can also be misused to create misleading content, so tech companies are developing ways to identify AI-generated material.
Large Language Models (LLMs)
Large Language Models (LLMs) use machine learning techniques to process and mimic human language. They are based on neural networks — computing systems inspired by the human brain — and are trained on massive amounts of text to learn patterns and relationships in language. LLMs can translate languages, answer questions in a chatbot format, summarize text, and generate written content such as stories, poems, and code. They do not have thoughts or feelings; they respond using learned patterns. Developers often fine-tune LLMs using a process called Reinforcement Learning from Human Feedback (RLHF) to make responses more conversational and accurate.
Machine Learning
Machine learning is a field of computer science, under the umbrella of AI, in which a computer system is taught to identify patterns and make predictions based on data. Data is repeatedly run through algorithms with different input and feedback to help the system learn and improve during the training process. Machine learning is especially useful for problems that are difficult to solve with traditional programming, such as recognizing images or translating languages. It requires large amounts of data, which has become increasingly available as more information has been digitized and computing hardware has grown more powerful.
Hallucinations / Fabrications
Generative AI systems can produce inaccurate responses that developers refer to as hallucinations or, more precisely, fabrications. Because these systems cannot distinguish between real and false information, they may confidently state incorrect facts or cite non-existent sources. Developers attempt to address this through a process called grounding, which provides the AI with additional information from trusted sources to improve accuracy on specific topics. Always verify AI-generated information before relying on it.
Responsible AI
Responsible AI refers to the practice of designing and using AI systems that are safe and fair at every level — including the machine learning model, the software, the user interface, and the rules and restrictions governing access. Because AI systems are created by humans and trained on data from an imperfect world, they can reflect inherent biases. Responsible AI involves understanding training data and finding ways to mitigate shortcomings to better reflect society at large, not just certain groups of people.
Prompts
A prompt is an instruction entered into an AI system — in the form of language, images, or code — that tells the AI what task to perform. Thoughtfully crafting prompts is essential to obtaining accurate, useful, and appropriate responses from AI tools. The more specific and detailed a prompt, the more likely the AI is to produce a relevant and helpful result.