Generative Artificial Intelligence (GenAI) or Large Language Models (LLMs) are tools that use deep learning models to produce new content such as text, visual content (images/video), and audio. They do this by analyzing large data sets, learning patterns and relationships, and generating new content that aligns with these patterns. It’s important to remember that Generative AI does not understand concepts and meanings as humans do, but instead imitates human behaviour based on probability.  

Generative AI is math, not magic. Generative AI works by learning patterns from massive amounts of data and then creating new content that follows those patterns. At its core, these systems use neural networks that process information through layers of interconnected nodes. During training, the AI analyzes examples (like text or images) and adjusts internal parameters to capture statistical relationships between elements. For large language models specifically, the system learns to predict what word should come next given previous words, developing a deep statistical understanding of language. When generating new content, the AI starts with some input (like a prompt) and uses probability calculations to determine the most appropriate next elements to add, creating a sequence that statistically resembles what it learned during training, without simply copying its training examples.

In algebraic terms, what LLMs do is recursive probabilistic inference. At each step they compute the probability of the next token based on all previous tokens, updating their internal probability distribution as more context is processed. Each generated token is chosen based on probability weighted continuations, continuously updating the model’s internal estimate of likely sequences. Mathematically this is considered ‘reasoning’ because it follows structured rules for updating likelihoods and making predictions…True cognitive reasoning involves conceptual leaps, independent hypothesis formation and genuine understanding.. none of which emerge from statistical inference alone. LLMs are not the foundation for true intelligence, AGI or whatever metaterm you prefer. The algebraic definition of reasoning is satisfied by current systems, but the richer cognitive definition remains beyond their reach. Denis O. GenAI models reason.. but there is a catch

Check for permission: Generative AI should only be used in a course, including tests or assignments, when an instructor has given explicit permission. Rules vary between courses, instructors and specific assignments, so check the instructions and, if unsure, discuss with the instructor. Using Generative AI when not explicitly permitted to do so can have serious academic consequences.    

Acknowledge the use of Generative AI:  When using Generative AI as part of an assessment, this should be clearly acknowledged. Never submit AI-generated content as your own work. Use the CapU Library Guide to AI citations in APA, in MLA formats. Keep notes as you’re working to track your usage.  

Be cautious and critical: AI-generated content may include inaccurate information and reflect biases from the data used to train the tool including linguistic, racial, gender and cultural biases. Fact check AI-generated content to ensure that it’s accurate and appropriate.  

Develop your own writing and communication skills: Use Generative AI to enhance your communication skills, not replace them. Don’t let it prevent you from developing the skills you need to succeed at and beyond university. Avoid using it for personal reflections or creative work that is meant to reflect your own unique perspective and ideas.  

Don’t input personal or sensitive information: Never input information such as your full name, birthday, address, health information or passwords into a Generative AI tool. Don’t enter copyright materials, such as resources from the Library or your instructor’s teaching resources. 

If you’re using Generative AI to complete activities or assignments for a course, you should refer to your course syllabus on the use of Generative AI and ask questions if the guidance is not specific. You should also review assignments for any particular instructions related to disclosure and citation of Generative AI use. If you’re unsure about what your instructor’s expectations are for disclosing the use of Generative AI, reach out for guidance. When in doubt, cite and disclose all uses of Generative AI. Use the CapU Library Guide to AI citations in APA, and in MLA formats. 

If the use of Generative AI is permitted in your course, consider exploring these uses:  

Create activities and problems to practice learning a skill or concept.

• Benefit: If you‘ve run out of practice problems in your course, this can be an easy, on-demand way to continue to practice learning a skill. 
• Risk: Depending on the discipline or topic, the solutions provided may not be accurate. There is also a chance that the problems and activities are not relevant or accurately framed, so you should be prepared to assess the quality of the output. 
• Additional tip: Consider including examples of activities and problems in your prompt to demonstrate the type of practice you are looking for. Input some feedback you received on an assignment and ask for exercises to help you build any missing skills. 

Request real-time feedback on your ideas

• Benefit: Generative AI may be able to provide counterarguments to a claim, identify holes in your logic, or identify opportunities to incorporate additional evidence to support a claim. 
• Risk: The feedback provided may be surface level or contradict the scope, goals, or parameters of a given prompt.beyond the scope of the request. 
• Additional tip: Use Generative AI feedback to help refine your ideas in the early stages of a project before seeking more in-depth feedback from peers, instructors, or other campus resources. 

Break down a task into multiple steps

• Benefit: When dealing with a difficult project or task, Ggenerative AI can help you break a task down into smaller, more manageable steps. For example, when preparing for an upcoming midterm or exam, Generative AI may be able to help you create a personalized study plan. 
• Risk: Generative AI may suggest unnecessary steps or omit important steps. 
• Additional tip: When you write your prompt, tell it what your final goal is and give it some examples of intermediate steps that you know you’ll need to follow. State both the goal and provide examples of intermediate steps you know need to be included. Using a prompt pattern like the recipe pattern may also help to improve the quality of the output. 

Generate paraphrased, summarized, and reformatted versions of content

• Benefit: When encountering complex, difficult, or lengthy content, Generative AI may be able to identify key takeaways or rephrase information in a way that is more accessible. 
• Risk: Taking the time to summarize material and put it in your own words can help with learning and comprehension. Depending on the learning context, using Generative AI to perform this work may prevent you from reaping the full benefits of this task. You may want to also consult with course materials, peers, and instructors to confirm that the information that you’re getting from a summary of a topic is accurate. 
• Additional tip: Consider prompting Generative AI to ask you questions about a given text and to provide feedback on your explanations of key concepts. You might also ask questions to the Generative AI platform in a way that mimics a conversation, which can be helpful when a tutor or an instructor is not available. 

(Adapted from Generative AI at Vanderbilt) 

It’s unclear how exactly Generative AI will impact the future of work. Some jobs will encourage or require Generative AI skills, while others will be more cautious or limited in their use. The pace of change is rapid, and the future is impossible to predict with certainty. In this changing landscape, you may want to seek out opportunities to learn more about how Generative AI tools really work, research how they’re currently impacting careers in your field and discuss this with instructors and other mentors.

When deciding whether to use Generative AI to complete a particular task, consider whether it will enhance your learning or diminish it. Ask yourself these questions: 

• Am I permitted to use Generative AI by this instructor? 

• If I use Generative AI for this task, will I still develop the skills that this activity is trying to help me build? 

• Am I knowledgeable enough on this subject to verify that everything is correct, accurate and free of bias and hallucinations? 

• Am I willing to take responsibility for any AI-generated content that I use?  

• Is this the best tool for the job or could I easily accomplish this with another, less environmentally costly tool? 

If the answer to any of the above questions is no, then it may be a sign that using Generative AI is not a useful strategy for this task.   

Check out the resources below and look for opportunities to learn more about how these tools work and how they are evolving. Take a deeper dive into some of the dilemmas related to intellectual property, environmental impacts and social impacts and consider what might mitigate or avoid those impacts in your field.  

LinkedIn Learning 

All CapU students have free access to LinkedIn Learning, a training platform with hundreds of short courses related to Generative AI including A Beginner’s Guide to AI, Generative AI for Business Leaders, and Ethics in the Age of Generative AI 

Coursera

Coursera offers a number of free courses on many subjects including Use Generative AI as Your Thought Partner, Generative AI for Everyone. Badges are available for a fee. 

Prompt Libraries

There are many prompt libraries out there that you can use and adapt for your purposes. This is a good one from Vanderbilt University: Prompt Patterns | Generative AI | Vanderbilt University

Student-Built Resources

There are a lot of resources out there – this one is built by and for students at the University of Sydney and includes lots of strategies for using it effectively: AI in Education