For most communities, another year of school has started, and one topic of conversation is what to do with AI. The current iteration of AI, generative AI, has exploded with services ranging from conversational bots to full-service school management platforms. In some ways, AI has followed the pattern of most new technologies in schools. The early adopters experiment with each tool and spread the good or bad word to the mainstream teaching population. However, generative AI is not simply the latest educational toy but a cultural tool/genre that could impact us as much as smartphones did 12 years ago.
The response from educators has been mixed; some schools are more inclined to restrict or ban AI tools, and others are finding good use cases that could support student learning. The responses are predictable based on the history of tech in schools. A fine line exists between AI deployed to support skills building and deeper thinking and AI deployed and inadvertently replacing those key learning elements. Some schools have scrambled to develop AI strategies and write policies before any significant use of AI tools in the classroom. It’s difficult to make any decisions when you don’t know the possible outcomes of an AI-supported program. Instead, schools must allow teachers to test drive some tools with their students and provide professional input. Then, schools can build a plan, not an AI plan, but an AI-supported learning plan.
Teachers’ experiences using AI tools with students will best inform policy decisions and raise questions regarding curriculum and pedagogy. Tech tools often do so, but because AI is potentially a broader and deeper support injection, the learning questions will be more challenging and numerous. One of the best examples of this process was published a few weeks ago in Edutopia Magazine. They interviewed Chanea Bond, a high school English teacher from Texas. The article is titled “Why I’m Banning Student AI Use This Year.” The title suggests that Ms. Bond is in the banning camp, but reading the article reveals a more nuanced set of circumstances. Ms. Bond’s fundamental argument is that her job is to teach students foundational skills, and AI is not a foundational skill. Her goal is for students to understand the literature they read and articulate their response to the reading with enough clarity that they can self-critique. Only then will they have the confidence to share their ideas.
Ms. Bond believes that generative AI preempts that deep thinking process by providing an interpretation and analysis of the reading that may or may not be accurate. Consequently, the student shifts from reflection and analysis to working with the AI bot’s explanation. Some argue that the evaluation and acceptance process is essential for students. Ms. Bond might agree, but she would likely say that it is not part of the current lesson, the goals of the course, or the state standards. One could question the efficacy of all three, which might be a good exercise for another discussion. Still, Ms. Bond is very clear about what she expects her students to learn, and AI tools do not, in her view, effectively support that process. In short, she may accept the potential value of AI in the classroom, but she believes that it should operate in support of skills that students have learned. From her vantage point, to do otherwise is to handicap students’ ability to transfer fundamental skills to future challenges. In this case, she does not want it to interfere with her students’ learning of the foundational skills of understanding and analyzing literature.
This case study’s message is how Ms. Bond decided to ban AI in her courses for the current year. In the previous year, she had the privilege of being allowed to fail, and that external vote of confidence motivated her to experiment with AI capabilities. Her decision to ban AI in her classroom did not re- sult from fear of the unknown or misunderstanding of the power of AI. She gave it a test drive and experimented with it but wanted better results. Students read a poem and were instructed to write a literary analysis. They had a choice of creating their thesis statement or asking AI to generate one, but if they chose AI, they were required to use whatever it came up with as an essay thesis. Ms. Bond reports that those who worked from AI-generated thesis statements wrote much weaker papers than those who developed their thesis independently. What were her reasons?
… my students don’t have the skills necessary to be able to take something they get from AI and make it into something worth reading. I also realized that they’re not using AI to enhance their work. They are using AI instead of using the skills they’re supposed to be practicing.
… my students didn’t know the poems they were feeding into AI well enough to tell right from wrong. And when they wrote their papers, they didn’t know the poem well enough, or their literary devices well enough, to take what they got from AI and make it their own.
She said that the students who came up with their thesis statements also wrote papers that needed some work, but their ideas were phenomenal. Learning to improve the mechanics of a written piece is a far easier task than learning the skills of critical analysis, interpretation, and argument develop- ment. The goal of the assignment was to develop personal ideas about the poem. She concluded, “To analyze your ideas, they must be your ideas in the first place.” Of course, the assignment could have been different. Students might have been asked to analyze another’s ideas, perhaps from AI. But that wasn’t the assignment. Ms. Bond’s experience was that AI obstructed rather than supported skills development.
The goal was the development of vital literary skills. Bond identified those skills as originality, crea- tivity, analysis, and synthesis. One could question the importance of those skills, and the resulting answers might impact the AI decision. For example, discussing students’ development of original ideas in teaching history has become commonplace over the past 20 years. As a retired history teacher, I rarely saw “original” ideas that were publishable (2 students in 27 years). My department repeatedly debated the meaning of original thinking. Was it a genuinely original idea developed by a high school student, or was it original to the student, meaning they had never thought about it previ- ously even though the idea was often well documented in historical scholarship (we tried to make the distinction between a high school history student and a graduate student of history)? The advent of the Internet and web searches pre-empted a resolution to the discussion. Most students simply did a Google search, read briefly about a few historical interpretations, and selected one to become the foundation of their argument. The quality of the essays suffered for the same reasons as Ms. Bond outlined in her literature classes, with less sophisticated technology. Perhaps this was not an AI issue but a more general issue of when tech tools genuinely help students and when they hinder their learning.
Literary interpretation and critique can be hampered by another challenge where AI can, on the sur- face, excel. Ms. Bond did not comment on the importance of background knowledge in supporting the development of fundamental skills. E.D. Hirsch argues that broad factual knowledge is essential for reading comprehension, critical thinking, and learning new information. He contends that stu- dents need more background knowledge to understand new concepts. How do we bring contextual background knowledge to the table in developing skills so students are more likely to retain that knowledge? There are several solutions: One is to use an AI bot because it can pull together back- ground knowledge more quickly and comprehensively than any human (of course, the results may not be entirely accurate, which is another skill for students to learn).
However, gathering large amounts of information and fantastic retrieval speed doesn’t make that knowledge stick. The time window for completing the assignment may not support adding that information to one’s experience set. There is no opportunity for repetition and transfer, but AI might help with those processes in the right environment. Another option would be to begin with content that captures the students’ attention (poetry, in Ms. Bond’s case). Hypothetically, what if Ms. Bond had given the students poems about climate change and global sustainability, a topic she knows they are passionate about (AI can identify those poems)? That passion brings with it background knowledge and a desire for more to complete the puzzle. Think about how a lack of background knowledge impacts your students’ work and how AI can make a difference in meeting student needs.
The Edutopia article presents an in-depth look at how AI creates challenges and questions for teachers and schools. On the learning side of the equation, the goal is to balance helping kids learn essential skills and providing support during that process, whether it be informational, reinforcing, and adaptive support or providing a personal learning pathway. There are no instant solutions, and your school/district will develop a unique approach based on reporting from teachers who experiment with the tools. Their job is to explain how to walk the fine line between the benefits of AI tools supporting learning and the problems of AI substituting for learning. There are materials to help, ranging from blogs to books. Tom Daccord’s AI Tools & Uses: A Practical Guide for Teachers is a comprehensive and targeted look at the AI tools that support effective learning, often with the added benefit of guardrails and safety. These tools allow teachers to spend more time with students and for students to spend more time reflecting on their work. Other tools give students a more personalized pathway to learning specific concepts and content. It would be hard to argue that those outcomes threaten or replace deep learning. They are your most valuable resource for application and fit. Then, you can develop strategies and policies aligned with culture and mission.
The productivity benefits of using AI tools (discussed above and in Tom’s book) may be an excellent first step to improving school and teacher productivity. Both the bots (ChatGPT, Claude, etc.) and some of the integrated school products (Magic School, SchoolAI, etc.) are capable of writing lesson plans, creating grading rubrics, generating and grading quizzes and tests, doing preliminary assessments of essays, leveling reading materials, and locating support materials without scrolling through pages of Google results. You can introduce these tools immediately because they will produce short-term results and a message to the teachers and students that AI will play a role in your school. It answers the calls for action positively and productively without the risks of immediately diving into support for student learning before teachers can work with AI learning tools. In the longer term, teachers’ careful examination of AI capabilities could result in conversations that impact curriculum development or pedagogy shifts. But most of us prefer to walk before we run.
About the author
Joel Backon is the editor-in-chief of Educate AI Magazine. Before joining Educate AI, Joel launched and was the founding editor of Intrepid Ed News, an online education publication that grew to almost 100,000 readers. For most of his career, he was an independent schoolteacher and administrator, teaching history and government while leading the school’s instructional technology effort, running a dormitory, coaching basketball, and chairing several committees. Immediately following college, Joel spent 15 years in the printing and publishing industries.
