AI basics for university teachers

As a teacher, how can I learn the basics of Generative AI?

Many faculty may be feeling frustrated that they need to learn about AI, especially if it isn’t relevant to their fields of study. However, when we look at the data about students’ AI use, we can see that there is an important role for us to play in helping our students navigate a new learning environment that includes access to many AI tools. The Digital Education Council Global AI Student Survey 2024 presents some useful data about the frequency and nature of students’ use of AI in educational contexts, as well as their attitudes about the role of AI in their learning in higher education. Below are just a few findings of this international survey of 3839 students:

  • 86% reported that they use AI in their studies

  • 54% reported that they use AI on at least a weekly basis

  • 58% do not feel that they have sufficient AI knowledge and skills

  • 52% believe that over-reliance on AI negatively impacts their academic performance

  • 86% reported that they are not fully aware of AI guidelines in their university

  • 73% agree that universities should provide training for faculty on the use of AI tools

These data make it clear that students are looking to us for help! Many of them are using AI frequently, but they realize that they are doing so in ways that are not well guided and that may even be harming their learning. At the same time, their desire for faculty who understand the use of AI tools shows that they realize we can’t help them without a basic understanding of AI tools, including their capabilities and their limitations.

This resource will provide an overview of generative AI and describe two ways you can experiment and gain basic familiarity with how it functions. 

What is Artificial Intelligence—and what is it not?

Artificial Intelligence is a broad term used to refer to the ability of a machine or computer to complete tasks in ways that simulate human learning. The AI tools that have received the most attention in the past two years represent a move forward in the technology and are called generative AI because they are able to create new content based on the data they have been trained on. 

While putting a prompt into a generative AI tool can quickly produce impressive results, it’s important to recognize that generative AI doesn’t actually think: instead, it learns patterns from the data it is trained on and uses those patterns to generate human-like text. Rather than producing anything original, it is basically mirroring the characteristics of the data it has seen before. Understanding this distinction is important because it can help us begin to recognize the limitations of what AI can produce—and to appreciate how that differs from what human intelligence can do. It can also help us think about the kind of thinking we want to make sure our students learn to do in our courses and make the case for why they need to do that thinking for themselves.

Getting started with generative AI

Learning to use generative AI is like learning to use most technologies: working hands-on is essential. To get started, decide which tool you want to use. Multiple generative AI tools are widely available, many in free versions that are less robust than paid versions. The good news is that while there are variations across these platforms, they all have the same basic functionality. If you just want to learn the basics of how generative AI works, Microsoft Copilot is a good starting point. All UAlbany faculty have access to this platform through Microsoft Office 365, and data in this system is protected because it is part of the University’s enterprise-wide license. Our licensed version of Copilot is also more robust than free versions of tools like ChatGPT or Gemini. You can access Copilot through the employee portal on MyUAlbany or by going to office.com and logging in with your UAlbany credentials. 

There’s an old expression in computer science: garbage in, garbage out. This means that when you put in poor quality information, you will get poor results. Prompting generative AI is a good example of this principle: the output you get will only be as good as the prompts you use. Experimenting with different kinds of prompts will help you get a sense of how the AI functions and start to make the kinds of adjustments that will be most productive for your specific goals. As you experiment, don’t just skip to the more detailed prompts—start with the more general ones so you can really see the differences.

Experiment 1: Use AI to get basic information

At its most basic level, generative AI can be used kind of like a sophisticated search engine where you look for answers to questions. Using it in this way is a good first step to gaining some comfort working within the tool and can show you what kind of output it can produce. 

Begin experimenting with this use of AI by prompting with a general question like this: “What is the best way to prune an apple tree?” When you pose a generic question like this, you will get a response that is probably accurate but is also likely to be very broad. 

Crafting a more detailed prompt will yield more useful information. You can provide context for your prompt, ask for specific types of output, and even specify what kind of sources you would like to get information from. For example, a better prompt for the apple tree question above would be: 

“I am a home gardener with a semi-dwarf Honeycrisp tree that is one year old. I don’t understand how to prune it and most of the information I get is confusing. Please provide me with guidance that includes a simple diagram of branches and shows me where to prune. I prefer information from a cooperative extension or from a university.” 

As you begin experimenting with AI, try putting in a general prompt and then a more detailed prompt about something that interests you. It is helpful to do some prompting about topics you know well so that you can get a sense of whether the output AI gives you is accurate and appropriately detailed.

Experiment 2: Use AI as a quizzing tool as your students might

For your second experiment with AI, we suggest you interact with AI as a student might. Most students continue to use ChatGPT when they do course work (Digital Education Council Global AI Student Survey, 2024). You can access ChatGPT without making an account at chatgpt.com.​ One of the ways that students can use AI to support their learning is for them to use it as a study tool. They can have AI quiz them so they can practice recalling and using key ideas and concepts. We suggest you try this for yourself. 

Begin experimenting with this use of AI by prompting with a question like this: “Can you quiz me on some basic questions related to _____?” Once again, you will probably notice that general prompts like this one don’t yield the most helpful self-quizzing questions.

Crafting a prompt that provides additional context and direction will yield better questions. Here’s an example of a more refined self-quizzing prompt: 

“I am a _____ level student taking a course in _____ and we have a test coming up on ____. Ask me questions about this topic in order to help me see what I understand and where I may need to study more. Ask me ONE question at a time and make each question short-answer or multiple choice. After I answer a question, give me feedback on my answer. Then ask me if I am ready for my next question. When I get a question wrong, make the next question a little easier. When I get a question right, make the next question a little harder. Keep asking me questions one by one until I ask to stop. When I ask you to stop, you should give me a summary of which parts of the topic I need to study more and also give me advice on the best ways to study what I need to learn based on research.“ 

You will likely notice that generative AI is much better at producing helpful self-quizzing questions with a very specific prompt like this one. 

Learning More about Generative AI

If you’re looking for more information about the different types of generative AI that are out there, we recommend checking out Ethan Mollick’s Substack newsletter, One Useful Thing. In “Doing stuff with AI”, a review from June 6, 2024, he gives an overview of many types of generative AI along with what they do well, what they do poorly, and what they cannot do at all.

For some thought-provoking conversations about how generative AI is impacting education, we recommend checking out Marc Watkins’ Substack newsletter, Rhetorica. 

References