Is ChatGPT good enough to take your job? Or is it simply another fad to ignore?
The internet is buzzing right now with opinions about ChatGPT—and its role in education—but there’s more to this generative AI tool than first meets the eye.
Many voices in the education sphere right now are thinking about how AI might affect learners directly—i.e. if students will use ChatGPT to cheat on essays or if students should use ChatGPT at all. Here at Ribbon, we’re curious—and cautiously optimistic—about how ChatGPT and other generative AI technologies might provide more support for teachers and other student-facing roles.
Can ChatGPT help us get better at offering students the support they need, at scale?
Our team has been experimenting and talking to other educators and leaders in student support about how ChatGPT might be used not to replace human educators and success coaches, but to support you—so that you can automate the repetitive parts of your work and better support learners.
Here's six ideas for how you can leverage ChatGPT for learner support.
Using ChatGPT to Scale Learner Support
ChatGPT, GPT-3 and other language generation models can be powerful tools for student-facing staff like instructors, advisors and program coordinators looking to provide their students with personalized, high-quality and scaled support.
How so? Consider these six ideas for combating burnout, reducing admin load and scaling student support through generative AI.
1. Creating personalized feedback on assignments.
Using ChatGPT, educators can automatically generate personalized feedback based on coursework, attendance and notes. Doing so could save time and allow you to provide more detailed and targeted feedback for each student. Of course, you don’t want to replace your own feedback with ChatGPT—but by allowing ChatGPT to generate some basic feedback, you get a head start rather than starting from a blank slate.
2. Generating additional study materials.
ChatGPT does an impressive job of creating practice problems, basic study guide outlines or summaries of key concepts. Generating these via ChatGPT can be especially useful for supporting students who need additional resources, while reducing teacher or advisor workloads.
3. Developing new template ideas.
Every great advisor and instructor has an arsenal of their greatest hit emails. By feeding your best email templates to ChatGPT and asking it to create additional templates for new or similar topics based on your current style and tone, it can generate new emails for you in no time.
4. Providing additional “office hours” support.
If you’re unable to meet with all of your students during office hours—as is common for online learning institutions—you could use ChatGPT to draft common student questions and provide additional support to learners who are unable to meet with you directly.
5. Preparing personalized lesson plans.
You could also use ChatGPT or other generative AI technology to help you create personalized learning plans for students based on their unique learning needs and goals. All of this data can be fed into AI tools to create more personalized responses. You’ll be able to ensure that students have the personalized support they need, but through a process that’s more scalable.
6. Supplemental feedback loops.
One of the best ways to cement what you’re learning is to teach it to someone else. But what about leveraging ChatGPT to develop student ideas and get 1-on-1 feedback? One faculty we talked, Prof. Jarrod Barnes from NYU, has been encouraging students to create ChatGPT prompts (allowing them to practice articulating their ideas more concretely) and use ChatGPT to get feedback on their ideas and concepts.
A Word of Caution: The Limitations of ChatGPT
For how powerful ChatGPT and other generative AI tools are—and promise to be in the future— it’s important to keep in mind their current limitations.
As Sam Altman, the CEO of OpenAI has written, “ChatGPT is incredibly limited, but good enough at some things to create a misleading impression of greatness. it’s a mistake to be relying on it for anything important right now. it’s a preview of progress; we have lots of work to do on robustness and truthfulness.”
As a result of the current limitations, it’s important to keep in mind that ChatGPT should be used as more of a muse, assistant or intern for your work, rather than a replacement for your knowledge and expertise.
Four current limitations of ChatGPT for educators to keep in mind:
It’s free (for now). ChatGPT is a machine learning model that learns from increased interactions with users. It’s beneficial for both demand generation and modeling to introduce ChatGPT to the general public for testing and continued development. However, it’s already assumed that the project will (at some point) have to be monetized. As a result, it’s important to keep in mind that ChatGPT, at some point, may not be as widely accessible.
It’s not always accurate. Although ChatGPT spits out answers nearly instantly, and with great confidence, the answers it gives aren’t always accurate. Since ChatGPT responds in each instance with confidence, it can be difficult, if not impossible, to spot factual errors unless you already know the correct answer. You should always carefully review AI outputs for accuracy.
Implicit bias/errors in ChatGPT. In some cases, ChatGPT’s answers may be completely false. However, in other cases, they may harbor more implicit bias, produce harmful suggestions or results, or contain erroneous conclusions. While OpenAI is working to eliminate these biases and errors, it’s still a work in progress.
Limited data sets and training. ChatGPT is trained on a massive, but still limited, data set. This in many cases leads to the two aforementioned problems around accuracy and bias. For example, ChatGPT’s training doesn’t include data from after 2021, making it’s knowledge of current events limited, and it can’t crawl the web for information, like Google.
All of this doesn’t mean we shouldn’t use ChatGPT in education, but that we should approach it wisely and be aware of its limitations. Rather than assuming we can “outsource” parts of our work we dislike to ChatGPT without a second thought, we should view generative AI technologies as assistants that can help brainstorm ideas, generate content based on a previous data set and reduce repetitive work and communications.
What do you think? Are you using ChatGPT in your work — or looking for ways to do so? Join our community of adult education operators in the Learner Success Guild. Through monthly events, our online forum, and more, you’ll find a space to share best practices, discuss challenges, and new ideas with other leaders from similar institutions.