Generative AI refers to algorithms that can generate new content, such as text, images, music, or code, based on the data they have been trained on. These systems, including popular models like GPT-4, can create human-like text, making them powerful tools for various applications.

Generative AI models are typically built using neural networks, especially deep learning techniques like transformers. These models learn patterns and structures from their training data. Here’s a simplified overview of how they work:

Training phase: Generative AI models are trained on large datasets, which could include books, articles, websites, and other text sources. The quality and diversity of the training data significantly impact the model’s performance and versatility. During training, the model learns to predict the next word in a sentence given the previous words, effectively learning the structure and patterns of the language.

Generation phase: Generative AI models start with a prompt, which is an initial input provided by the user. This could be a word, a sentence, or a more complex instruction. When given a prompt, the model uses its learned knowledge to generate new content that follows the patterns it has seen during training.

Common types of generative AI models include the following:

• Language models/LLMs: These models, such as GPT-4, are trained on vast amounts of text data and can generate coherent and contextually relevant text. They can write essays, answer questions, summarize information, and even engage in conversations.
• Image generators: Models like DALL-E create images from textual descriptions, allowing users to generate visual content from scratch.
• Music and audio generators: These models can compose music or generate sound effects based on given parameters.
• Code generators: AI tools like GitHub Copilot assist in writing code by suggesting snippets and completing code blocks based on the context provided by the user.

Since OpenAI’s release of ChatGPT 3.5 in November 2022, generative AI (ChatGPT in particular) has quickly become a hot topic in higher education, especially in learning assessment and academic integrity. Regardless of how you intend generative AI to be used (or not) in your courses, we recommend that you try out some generative AI tools, such as ChatGPT (OpenAI), CoPilot (Microsoft), Gemini (Google), etc., to get a feel for what AI can and cannot do. This will help you to make informed decisions on whether and how you will use or allow the responsible use of AI in your classes.

Resources to get started

Check out this AI Guide by Harvard AI Pedagogy Project. Part I AI Starter and Part II LLM Tutorial will help you get started with the generative AI tools. Alternatively, take this 2-hour course Innovative Teaching with ChatGPT (by Vanderbilt) on Coursera.

Students might use generative AI for various academic tasks, which can be both beneficial and potentially problematic depending on the context and extent of use.

• Academic writing and essays
• Research and information gathering
• Study assistance and exam preparation
• Coding and prolem solving
• Creative projects
• Language learning and translation

Generative AI offers powerful tools that can significantly enhance the learning experience when used ethically and responsibly. However, it also poses risks to academic integrity if misused. By understanding how students might use these technologies, instructors can design assignments that encourage productive use, foster deep learning, and uphold the values of academic honesty.

Yes! The University provides a syllabus statement guide for AI with examples for classes where AI is allowed, classes where AI is partially allowed, and where AI is not allowed.