Nav Screen Region
Site Header Region
Page Header
Main Content Region

Introduction to Generative AI

Overview

Artificial Intelligence (AI) is revolutionizing our world in ways that higher education is only beginning to comprehend. As institutions of learning, it is imperative that we - educators, researchers, instructional designers, and students - diligently work to understand this rapidly advancing technology. Higher education's role is to equip all members of the academic community to both critically examine and judiciously apply emerging technologies like AI. We must learn what AI can do, how it is being developed and regulated, and how it might transform teaching and learning.

Generative Artificial Intelligence (GenAI) enables the creation of novel content derived from a user's inputs, whether textual, visual, auditory, or video-based prompts. OpenAI’s Conversation Model ChatGPT, released in late 2022, demonstrated generative AI's widespread appeal by amassing millions of users within a short time. Concurrently, other organizations continue advancing generative tools at a rapid pace to realize artificial intelligence's estimated $1.3 trillion market potential within the next decade. As AI's influence grows exponentially, higher education must ensure educators and learners develop AI literacy to navigate this shifting landscape.

Developing AI literacy will be essential for both instructors and students. For students and later graduates will inevitably encounter and employ AI throughout their careers and continued studies. Becoming AI literate will support students’ efforts to critically engage with AI systems’ outputs, be aware of potential biases, and develop essential critical thinking and problem-solving skills. For instructors, understanding (and playing with) AI should enable us, where appropriate and applicable, to better use AI tools in our teaching to improve the overall learning experience for students. By learning AI's capabilities, limitations, and ethical hazards, instructors can make informed decisions on how, if they should choose, to incorporate AI into their courses.

Generative AI remains a dynamic field where academia has only begun exploring applications for teaching and learning. This resource will serve as a starting point for ongoing dialog and experimentation around AI's instructional roles. As the technology and our understanding mature, this page will be revised to track emerging practices, opportunities, and considerations regarding generative AI's integration into curriculum development, pedagogy, and assessment. If you would like to talk about GenAI and your concerns or ideas, please feel free to contact us at FCE@csusb.edu.

AI Literacy

The resources below will introduce you to foundational knowledge about artificial intelligence, AI's role in education, and the generative AI landscape. This resources are designed to help you develop your own AI literacy and your ability to thoughtfully integrate AI into teaching practices and learning experiences.

Glossary

This Glossary of AI terms will help you become more familiar with AI-related terminology.

What is Artificial Intelligence (AI)? 

AI has a range of definitions and a spectrum of degrees depending on the context. As a science, artificial intelligence can be described as combining “the three disciplines of math, computer science and cognitive science to mimic human behavior through various technologies”.

As defined by IBM, “At its simplest form, artificial intelligence is a field, which combines computer science and robust datasets, to enable problem-solving. It also encompasses sub-fields of machine learning and deep learning, which are frequently mentioned in conjunction with artificial intelligence. These disciplines are comprised of AI algorithms which seek to create expert systems which make predictions or classifications based on input data.”  In Oxford Language, “The theory and development of computer systems able to perform tasks that normally require human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages.

 

Types of Artificial Intelligence (AI):

There are different types of AI, including:

  1. Narrow AI: Also known as weak AI, narrow AI is designed to perform specific tasks within a limited domain. Examples include voice assistants like Siri and Alexa, recommendation systems, and image recognition algorithms.
  2. General AI: General AI refers to AI systems that possess the ability to understand, learn, and apply knowledge across different domains, similar to human intelligence. However, true general AI is still a theoretical concept and has not been fully realized.
  3. Superintelligence: Superintelligence refers to AI systems that surpass human intelligence in virtually every aspect. This level of AI is purely speculative and is the subject of much debate and discussion among researchers and experts.

Check out these videos:

  1. Technology with AI at its heart has the power to change the world. The Royal Society.
    @What is artificial intelligence? | The Royal Society
  2.   Master Inventor Martin Keen explains different AI categories. IBM Cloud
    @The 7 Types of AI - And Why We Talk (Mostly) About 3 of Them
What is Generative AI (GenAI)? 

Generative AI is a type of artificial intelligence technology that can learn from existing data to generate new, realistic content that reflects the characteristics of the training data but doesn't simply repeat it. It can produce various types of content, such as images, video, music, speech, text, software code, and product designs. It works by deriving patterns from large sets of training data that become encoded into predictive mathematical models, a process commonly referred to as ‘learning’. Generative AI models do not keep a copy of the data they were trained on, but rather generate novel content entirely from the patterns they encode. People can then use interfaces like ChatGPT, Claude, Bard, DALL-E 2, or MidJourney to input prompts – typically instructions in plain language – to make generative AI models produce new content. Expanded info can be found at Techtarget by George Lawton, May 2023. 

 

Generative AI in a Nutshell

This video shows how to survie and thrive in the age of AI. It covers questions such as: 

  • What is generative AI? 
  • How does generative AI work? 
  • How do I use generative AI? 
  • What are some the risks and limitations of generative AI?

The video also covers autonomous agents, the role of humans, prompt engineering tips, Ai-powered product development, the origin of ChatGPT, different types of models, and some tips for coming to terms with generative AI.

How AI Chatbots Work

Large language models (LLMs) are the fundamental architecture behind chatbots like ChatGPT or Bard. A question typed in to ChatGPT, such as “What is the capital of France”, has to be processed by an LLM in order to produce an answer like “The capital of France is Paris”. Here’s a visual walk-through of how this type of artificial intelligence works.
  How AI chatbots like ChatGPT or Bard work – visual explainer by Seán Clarke, Dan Milmo and Garry Blight, The Guardian, November 1, 2023

Generative AI Tools

Generative AI tools such as ChatGPT, Bing, Bard and Claude operate by responding to user-provided prompts or queries. The process of carefully crafting an effective prompt, also called prompt engineering, is integral to eliciting high-quality responses from language models. While some tools focus on text generation, others are capable of producing novel images, audio, video or computer code based on the user input. Prompt creation requires strategic phrasing to maximize the utility of these generative tools across various media and applications. By refining the initial prompt, individuals can steer the AI tool's output toward more insightful and appropriate responses. Therefore, crafting prompts represents a key skill for fully realizing the potential of these advanced artificial intelligence technologies.
 

ChatGPT “is a variant of the GPT (Generative Pre-training Transformer) language model, which was developed by OpenAI. GPT models are trained to generate human-like text by predicting the next word in a sequence based on the words that come before it. ChatGPT is specifically designed to be used in chatbots and conversational systems, and it is trained on a large dataset of human conversations to learn how to generate appropriate responses in a variety of contexts.

  

  • GPT-4: the newest version of OpenAI’s language model systems, officially launched in March 2023. It is a multimodal model, in that it accepts text and images as input. It can only generate text. It has demonstrated much stronger academic performance compared to the GPT-3.5 model. Learn more: OpenAI’s launch announcement on GPT-4.
    • The ChatGPT-4 tool based on this model is only available through paid ChatGPT+ subscription at this point ($20/month.)
  • Bing: Microsoft’s AI chatbot. After the release of GPT-4, Microsoft officially confirmed that Bing runs on OpenAI’s GPT-4 model.
  • Dall-E 2: ChatGPT’s visual creation sister, also run by OpenAI.
  • Bard: Google’s AI chatbot was updated with the Gemini large language model in December 2023. Both are based on Pathways Language Model 2 (PaLM 2) and Google's Language Model for Dialogue Applications (LaMDA)
  • Claude: Claude is a powerful AI model developed by Anthropic that can process and understand large amounts of text, enabling various applications such as summarization, Q&A, forecasting.

 

What is Generative AI and How Does it Work?

An interactive lecture by Mirella Lapata, The Royal Institution (Oct 12, 2023)

ChatGPT 4.0 generated summary of the video: 

  1. Definition and Scope: Mirella Lapata defines generative AI as a computer program creating new content (audio, code, images, text, video) using partial existing information. The lecture focuses on text generation due to Lapata's expertise in natural language processing.
  2. Historical Context: Generative AI isn't new; examples include Google Translate (since 2006) and Siri (since 2011). These are early forms of generative AI, demonstrating its integration into everyday technology.
  3. Technology Behind Generative AI: The technology relies on language modeling, predicting the likelihood of the next word in a sequence. This involves using neural networks and a process called self-supervised learning, where a model is trained to predict parts of text that are intentionally omitted.
  4. Development and Challenges: Generative AI has evolved significantly with models like GPT-4, which require large datasets and computational power. Key challenges include managing the model's alignment with human values, avoiding biases, and ensuring accurate and non-offensive outputs.
  5. Future Prospects and Regulation: The future of generative AI is both promising and challenging, with potential benefits and risks. Regulation is anticipated, similar to other powerful technologies like nuclear energy, to minimize misuse and maximize benefits.
Generative AI Detection Tools

AI detection tools, like that in Turnitin or GPTZero, are controversial due to their unreliability. Instructors should be aware that false positives and negatives are possible and frequent with such detection tools. Because language models generalize and summarize existing knowledge based on probability predictions of word sequences, rather than copying it verbatim, it may be impossible to identify when GenAI has been used with certainty.

No software is able to detect AI-generated text with 100% certainty, and multiple false positives may seed mistrust between students and instructors, compromise access needs, and amplify systemic inequities in higher education.

More information about testing of these tools, including Turnitin, can be found in these reports from Tech Crunch (2023, February 16), Inside Higher Ed (2023, June 1), and Bloomberg(2023, September 21).

 

Review this interactive lecture by Mirella Lapata: What is generative AI and how does it work? The Royal Institution (Oct 12, 2023)

 

ChatGPT 4.0 generated summary of the video: 

  1. Definition and Scope: Mirella Lapata defines generative AI as a computer program creating new content (audio, code, images, text, video) using partial existing information. The lecture focuses on text generation due to Lapata's expertise in natural language processing.
  2. Historical Context: Generative AI isn't new; examples include Google Translate (since 2006) and Siri (since 2011). These are early forms of generative AI, demonstrating its integration into everyday technology.
  3. Technology Behind Generative AI: The technology relies on language modeling, predicting the likelihood of the next word in a sequence. This involves using neural networks and a process called self-supervised learning, where a model is trained to predict parts of text that are intentionally omitted.
  4. Development and Challenges: Generative AI has evolved significantly with models like GPT-4, which require large datasets and computational power. Key challenges include managing the model's alignment with human values, avoiding biases, and ensuring accurate and non-offensive outputs.
  5. Future Prospects and Regulation: The future of generative AI is both promising and challenging, with potential benefits and risks. Regulation is anticipated, similar to other powerful technologies like nuclear energy, to minimize misuse and maximize benefits.
Additional Resources--Learn More