Last updated on September 4, 2024
It seems like higher education has been through a roller coaster over the past several years. The challenges of the pandemic, an explosion in online enrollments, and the transition back to the classroom may have left many educators and students dazed and trying to adapt. Just when there was a chance to catch our breath, an emerging technology asserted itself with the power to completely revolutionize teaching and learning: generative artificial intelligence, or GenAI for short.
ChatGPT, the first mainstream Large Language Model (or LLM), and competing tools are enthusiastically pushing the boundaries of what technology can offer, not least of all in education. Part of this can lead to unsettling questions about academic integrity and how we should approach the assessment of learning. On the other hand, this radically new and evolving technology can be leveraged to improve the learning experiences we offer to students. Going a step further, we now see the need to prepare both faculty and students for a future that includes this disruptive technology is both uncertain and also could be very rewarding.
As a brief introduction, an LLM is an algorithmic computer program that can analyze patterns in written language and then generate brand-new compositions in response to a user-provided prompt. They have myriad uses, like researching unfamiliar topics or developing written works. There are even image generators that work in much the same way. Like any new educational technology, there is the potential for misuse and abuse, but introducing your students to its proper role in their education is the best way to encourage them to use it to enhance their independent thought rather than replace it.
In this spotlight, we’ll look at some ideas for how generative AI can be used in education, including some specific uses for MSU Denver instructors. A sizable number of MSU Denver faculty are already using GenAI as a tool in their instruction, but many others report hesitancy due to such factors as unfamiliarity and ethical concerns. Potential abuses are not trivial, but here we will focus on the possible advantages you can gain by implementing this technology. Generative AI is still in its infancy as a consumer product, and there will be much more to learn as it matures. CTLD is committed to being a resource to MSU Denver educators who are eager to explore its potential.
Contents
- Use Microsoft Copilot – this tutorial covers the basic use of Copilot, like entering prompts and using the interface.
- Copilot Example Showcase – this showcase is a collection of practical demonstrations of using Copilot.
Let’s walk through it together
Best Practices
How does generative AI work?
- GenAI is any kind of computer program that can produce novel creative works by analyzing and mimicking existing human-created work, like writings or images. Large language models, like the famous ChatGPT, are a kind of GenAI focused on writing. They essentially operate by predicting what words could follow another in a sentence based on patterns it observes in human language.
- The product of generative AI can only ever be as good as the material it has analyzed; it isn’t capable of independent logic and reasoning. For that reason, it can easily parrot misinformation or conform to cognitive biases already present in human writings. It can also hallucinate (an industry term) new information, as its predictive algorithm may lead it to combine ideas in a false or misleading way.
- Despite these flaws, it is still an incredibly potent way for information to be parsed and presented in an easily comprehensible summary. Think of it like a kind of search engine: Google can link to websites with true, false, or questionable information. It cannot make judgments about what is correct, only about what is popular.
- Large language models are capable of responding to a huge variety of prompts. It can answer questions, follow directions in what it generates, and refine the output based on user input. Users are free to communicate with it conversationally, clarifying their requests or asking follow-up questions in natural language.
How could this be used in education?
- While GenAI is similar to a search engine, it is far more powerful. When searching the internet for information, the user still needs to follow links, read a lot of information, decide what is relevant and synthesize the results. GenAI can absorb and condense a great deal of information quickly before presenting it to the user.
- GenAI can also analyze writing the user provides, offering advice on the strengths and weaknesses of what it reads. It could be used as a tutor or coach for writing by students. Of course, it can only compare the writing to the data set it has available to it and can’t be relied upon to give expert opinion.
- You and your students can use generative AI conversationally, asking questions and using it to brainstorm new ideas. In this role, it is a kind of collaborator, offering feedback and allowing you to express your ideas. It isn’t a replacement for a trusted colleague, but it is always available for you to use.
- It is also possible for anyone to try to pass off generated content as their own writing. While this is a real risk, students who know how to use GenAI as a research assistant, coach, or collaborator may be less likely to use it as a replacement for their own intellect.
- GenAI is an emerging technology, and while some attempts have been made to develop ways to detect content that has been created using it, these detectors have troubling rates of false positives. We caution you against relying on tools like these and instead focus on educating students on using the tools responsibly rather than abstaining from them completely.
How do I use this in my course?
- The single best (and easiest) thing you can do is communicate your policy regarding generative AI with your students immediately. Let students know what kind of use is and isn’t acceptable in your course. Make recommendations on how they might use it most effectively. If there’s a specific research project you think it might be helpful for, mention that. At the same time, let them know that they still need to cite peer-reviewed research and they can’t submit AI-generated work as their own.
- Try out GenAI for yourself to understand its abilities and limitations. All MSU Denver employees have access to Microsoft Copilot. Copilot is preferred for professional use since MSU Denver accounts have their data protected. You won’t need to worry about potentially sensitive data being saved by the program, especially that protected by FERPA. A good place to start is to just ask it how Generative AI can be used for your subject matter.
- Create assignments that instruct students on using it appropriately for your class. For example, you can instruct students to start their research with MS Copilot and then compare the results to resources from the Auraria Library. Have students report how accurate their results from Copilot were and what things it might have missed in the peer-reviewed research they found.
Further Resources
- The Cornell Center for Teaching Innovation has a series of articles on the role of generative AI in education.