AI – The Promise and the Peril: Managing AI in the Classroom
By Marilyn Pongracz
AI is everywhere, from a toddler companion toy (Rosenberg, 2023) to prosthetic hands that can adapt to the user (Arhirova, 2023). AI has been developing for years, but generally was unnoticed until ChatGPT was introduced in November 2022.
Many have expressed fear that the reach of this technology will overtake human beings. Noam Chomsky addresses this. “Today our supposedly revolutionary advancements in artificial intelligence are indeed cause for both concern and optimism. Optimism because intelligence is the means by which we solve problems. Concern because we fear that the most popular and fashionable strain of A.I. — machine learning — will degrade our science and debase our ethics by incorporating into our technology a fundamentally flawed conception of language and knowledge.” (Chomsky, et al., 2023).
The article then describes significant differences between human reasoning and machine learning. AI systems pull vast amounts of information and data, find patterns in that data, and combine what is found to generate output that mimics what human beings can do. However, although people can only manage a little data at a time, the learning process is quite different. First, with only a relatively small amount of input, children instinctively acquire language. Second, machines cannot reflect on the past of what might have happened nor can they predict the future of what could occur unless given explicit instructions of the conditions for those events. Third, while human beings make mistakes, they know they may be wrong and can actively seek correction. Finally, as evidenced in a conversation about the morality of inhabiting Mars, a researcher demonstrated a lack of morality in AI. It produces both what is true and what is false without regard for the consequences of its output. (Chomsky, et al., 2023)
2023 Spring Conference Silver Sponsor
For educators, AI has made some of what we do easier, for example, differentiating texts or creating assignments. In fact, with some expertise, it can be taken further. One educator “employed Botsonic [an AI chatbot that can be trained on any customized data] to develop AI software that can align learning objectives or specific components of a lesson or initiative with the [state’s] standards” (K. Ware, personal communication, January 3, 2024).
The consensus is that AI cannot be ignored, but to this point, educators are struggling to find ways to make use of it to assist rather than obviate learning.
One creative option that some teachers have tried is to ask the AI to impersonate a famous person or someone from history, have students ask it questions, and then compare their answers. Another is to start a student-generated story, then ask the AI to suggest 3-4 options for the next part of the story, and finally in groups, students can follow the different prompts to generate different stories that they can share with their classmates. (Peachy, 2023) This author has developed numerous prompts in an online book for purchase.
Another professor, Leah Carmona, used AI to encourage her high-beginner students to understand the importance of word studies. They use various online dictionaries such as the free Longman Dictionary of Contemporary English Online in reading classes to find synonyms, antonyms, and related words.
When previewing “pet,” students learn collocations, such as to have a pet or make good pets, and related words like domestic or exotic. They also create a word wall to brainstorm adjectives and ideas to describe their favorite pet. After that, students were asked to write a short paragraph using some words discussed in class and submit their writing in Canvas. Unfortunately, sometimes students do not make an effort or do not find this exercise effective, so to help them be more creative and visualize vocabulary, she enlisted the help of DALL-E to generate pictures that could illustrate their vocabulary and understand what the reader would ultimately see.
The prompt was straightforward: “Generate an image for “student’s description.” Those with few adjectives or related vocabulary saw a simple and sometimes unclear image. However, the students who applied themselves had results like these:
“My pet is my dog, Max. He is a loyal and active friend. Max is a medium-sized dog with soft brown fur and dark brown eyes full of love. Max is there every morning when I wake up, ready to start his activities.”
They were amazed at what this innovative app was able to produce.
For another assignment, she used DALL-e to help illustrate their topic and concluding sentences so they could see how descriptive they were.
One student wrote: “Generate an image for Topic Sentence: “Last summer, my family and I went to the beach, and it was the best vacation ever.” and Conclusion Sentence: “When we returned home, I was smiling because I was thankful for the beautiful memories.”
Overall, this was very effective because it encouraged students to pay attention to their choice of words and helped them understand the importance of word study. (L. Carmona, personal communication, January 2, 2024).
Although there are some creative options, AI has challenged traditional assignments. Students can seem to produce amazing essays. One of the students in my advanced ESL writing class handed in an essay about a life lesson with this introductory paragraph.
“Life is an intricate journey, and woven into its fabric are lessons that often emerge from unexpected sources. One profound lesson that has profoundly shaped my outlook is the power of resilience, a gift from my grandmother. Through her unwavering spirit, she not only taught me to weather storms but also revealed the transformative strength hidden within life’s adversities.“
Obviously, the vocabulary and syntax are far more sophisticated than what the student could produce, and the paragraph lacks personal references.
An early possible idea for assisting learning was to have the AI app, ChatGPT, generate ideas to help students figure out what to write, but few would disagree that it is better if students learn how to brainstorm on their own.
Another solution is to create assignments that require students to compare AI output with their work, and then ask ChatGPT for more specific information to demonstrate its shortcomings. This involves critical discussions about AI content, which might be difficult for MLs who are still acquiring the language.
(See more ideas below.)
