Artificial Intelligence: Chat Bots and Language Models in English Language Arts
Are We Doomed?
By Hannah Lewis
This month, I had the incredible opportunity to participate in WPCTE’s HOPE through JOY conference at Penn West University in California, PA. This conference brought together educators from all over the region, and I was delighted to hear from teachers in such a wide variety of settings and contexts.Regardless of our contexts, one issue impacting all of us in education (and, honestly, in a variety of other fields) is the role artificial intelligence technologies–language models/chatbots like ChatGPT in particular–are going to play in our lives, careers, and classrooms.
While I came prepared to demystify these technologies and think through what they mean for our classrooms, I was delighted to see that another group of presenters from Grove City College (Fecich et. al.) had come to talk about the same topic. Their presentation and resources (linked here and available in the conference resources) solidified for me the relevance of the topic to educators at a variety of levels.
This topic began to interest me in December 2022 when I noticed that many of my students’ essays about Edgar Allan Poe’s role in American Romanticism contained similar features and all seemed to reference Arthur Conan Doyle. The culprit, I quickly discovered, was ChatGPT.
While my initial reaction to plagiarism is always frustration and a little reflexive anger, I talked with my 11th-grade English colleagues and we slowly began to see this as a learning experience and to embark on a journey of exploration together.
Our journey went something like this:
We were terrified that students would never write a paper themselves again.
We were arrogant enough to think we could create lessons to evade the problem– “ChatGPT can’t include direct text evidence and citations–we’ll just have to put that on all our rubrics now!”
We were at once both highly critical of and highly impressed by the essays ChatGPT could produce.
We eventually landed in a healthy place, where we established that, indeed, these types of A.I. are going to be with us for the rest of our careers, and we have to find a way to coexist with them.
That place of eventual, tenuous peace is the place from which I created my presentation, with a tremendous amount of help from my colleague and friend Abby Weller-Hall and acquaintance and applied data scientist Anthony Olivieri.
My presentation can be found here, with links to activities and resources, but the key takeaways can be found below:
- Chatbots are not actually intelligent or learning. Indeed, they are highly sophisticated and impressive pieces of technology and can (as Fecich et. al. describe) serve as very helpful productivity tools for educators and other professionals. However, they cannot think critically or creatively and, therefore, cannot replicate the most important skills we aim to teach in the English Language Arts classroom.
- Plagiarism has been around forever and will always be a problem, as long as students are incentivized to work for something other than the experience of learning (grades, accolades, acceptance from teachers or peers, etc.). While tools exist to help spot cheating (like this excellent resource from CommonLit and Quill that not only provides a chatbot plagiarism “detector” but also a guide for how to approach the issue), we will always miss something. We cannot expect ourselves to perfectly catch cheating every time it happens. Instead, we should focus on creating assignments that are meaningful and engaging and convince students of their value so that, to the extent possible, they are motivated to do the assignments for the inherent value they have, not just for the grade.
- Similarly, technology continually inspires fear. Socrates resisted the written word, math teachers resisted calculators, and English teachers resisted spelling and grammar checkers, but now all of these tools have improved teaching and learning, allowing us to move beyond lower Depth of Knowledge skills and teach higher-level thinking skills. The same will be true for ChatGPT and other language models–we just have to figure out how to coexist with and, eventually, leverage them, just as we have every other piece of technology.
- While these tools are very good at writing unimaginative, template-like essays, they consistently fail to go deeper and, for example, write engaging introductions or thoughtful conclusions. This is our chance to move away from teaching boring, formulaic writing structures like the five-paragraph essay. Instead, we can take this opportunity to invite students to engage with how books connect to their lives and the implications they have on the world around them. Not only will reimagining writing in this way engage students more and make them less likely to want to plagiarize–they will simultaneously make plagiarism more challenging, since algorithms do not have lives or experiences to tie these pieces into.
While I’m still working to find ways to engage with ChatGPT and other language models, you can find some of Abby's and my attempts to leverage ChatGPT essays to teach critical thinking in writing here and here, respectively.
Let’s not sound like fearmongers or Luddites. Let’s find a way to live with and, ultimately, leverage these technologies to make our lives easier and teaching and learning better!
To see more of the great ideas from educators around the region, check out the conference website here and be sure to join us in 2024!
No comments:
Post a Comment