In which I respond thoughtfully to a journalist’s question about AI in schools, watch my nuanced argument get reduced to a single quote, and reflect on how complexity gets flattened at multiple levels—from educational policy to media coverage.
I recently wrote a post titled ‘F*** Nuance‘ about how academics often get bogged down in unnecessary complexity. So it’s ironic that here I am, just a week later, bemoaning about situations when nuance gets lost. I guess my ability to have it both ways is getting to be a habit. Which, as it turns out, was the theme of another post I wrote last week.
Be that as it may, I want, in this post, to focus on another tension, that of how educational complexity gets flattened in most conversations about the role of AI in education. Sometimes it is due to institutional policies seeking clear rules or by media coverage needing digestible quotes. Last week, I experienced both firsthand when I received an email from a reporter at the New York Post, asking for my thoughts on a story she was working on. Townsend Harris High School in Queens had decided to require students to handwrite their summer reading essays in class, abandoning their traditional take-home typing policy. The reason? To combat the widespread use of AI tools like ChatGPT that had made it nearly impossible to ensure authentic student work. Her questions were specific:
Whether policies like this are appropriate or overly reactive; How schools should be addressing AI use in student writing; Whether in-class handwriting offers any long-term solution or is a stopgap.
I spent some time crafting a thoughtful response, trying to capture the contextual nature of good educational decision-making. The story did make it to print, but as is natural in journalism, my extensive thoughts were condensed into one pithy statement. Since I believe this topic deserves the fuller consideration, and because I did spend the time writing out my complete perspective, I’m sharing my entire response here.
My Full Response to the Reporter
Thank you for reaching out with your question.
It is not surprising to me that individual schools are setting their own rules, given the lack of a district wide policy. This tech is so new, has spread so fast that institutions are unsure as to how to move forward.
These handwriting policies are understandably reactive – and there’s nothing inherently wrong with that response given how quickly AI has emerged. However, blanket policies become problematic because they prevent students from learning to work critically with these tools. [NOTE: A comment on LinkedIn forced me to revisit the previous statement. I think I missed the point there. What I was trying to say was that “Blanket policies become problematic because they impose uniform solutions on inherently diverse learners. Students don’t come to us as standardized units – they have jagged profiles of varying talents, interests, and knowledge that mean some students might benefit from AI assistance while others need to develop certain foundational skills before they’re ready to use AI effectively.] What GenAI has done is brought into question some fundamental assumptions about how we evaluate learning. There are contexts where having students struggle with their ideas, to craft a coherent piece of writing is critical. In these contexts, having students write (by hand or on a computer with no access to AI) does make sense. There is a necessary friction in learning that just having a chatbot complete the assignment sidesteps, and this handwriting approach responds to that concern.
But blanket policies are always concerning to me – just given the complex, context-driven nature of education. For instance, there are other situations or approaches where using a chatbot or AI may be useful. Additionally, AI can serve as a debate partner, helping students practice argumentation and consider counterpoints. It can suggest new approaches to writing, offer different rhetorical moves that students can then adapt and make their own. In science, students can create complex simulations using just plain English descriptions – that can be incredibly powerful for learning by creating.
But more importantly, students need to learn how to work with these tools given that they are now embedded everywhere. As we know, these systems are not perfect – they are overly sycophantic, prone to errors, biased, and often produce plausible-sounding but incorrect information. These limitations could be valuable teachable moments.
A blanket ban eliminates these learning opportunities entirely.
So as in all matters related to education these are complicated issues, depending on instructional goals, content, age level and more. As for handwriting specifically, I don’t see it as merely a stopgap. There will always be contexts where we want students to think through ideas without technological assistance – just as there are times when collaboration with AI tools makes sense. The key is matching the method to the learning goals.
In brief, I don’t think blanket policies are the solution. And honestly if the district is to come up with a policy it should be about giving freedom to schools and educators to play and experiment. In certain cases, using AI is the way to go, in others having students write and debate without AI assistance may be the right choice.
Now, you were most probably looking for a pithy quote, and I, as professors tend to do, have gone on and on. You can of course use anything in this message, but if I had to summarize this, here’s what I would say:
The real issue isn’t whether AI is good or bad for learning – it can be both depending on context. It is whether we trust educators to design appropriate learning experiences for their classroom, or impose blanket policies that strip away their professional judgment.
~ punya
After all this, what ended up in the New York Post story (Top Queens high school bans summer do-at-home essays to curb ChatGPT cheating) was a single quote:
Punya Mishra, director of the Learning Futures Institute at Arizona State University, said Townsend should be focusing on bringing AI into the classroom, rather than locking it out.
“Handwriting essays in class can make sense in certain cases,” Mishra said. “But blanket policies like this take away educators’ ability to teach students how to work with AI critically and wisely.”
The quote they used wasn’t even the pithy summary I had provided at the end of my email, but rather a sentence pulled from the middle of my longer response. And I understand why this is the case. It is simply how journalism works, editors need concise, quotable material that serves the story’s flow. That said, what was lost in the condensed quote was my emphasis on two key ideas: the agency of educators to make context-appropriate decisions, and the fundamental importance of matching educational methods to specific learning goals rather than applying one-size-fits-all solutions.
A Postscript
The story ended up being syndicated far more widely than I expected, particularly across Indian media outlets. It ended up in India Today, Business Today, Times of India, Pragativadi, and NDTV. It’s fascinating how a story about a Queens high school can spread globally, or at least to India 🙂
One amusing detail: while the New York Post managed to get my institutional title wrong in their story, most of the Indian publications just went with my broader institutional affiliation (Arizona State University). It’s a small reminder of how information travels and transforms as it moves through different media ecosystems.
In the grand scale of things, that my quote was simplified isn’t particularly significant—this is how media naturally works. But in educational contexts, this same flattening of complexity becomes much more consequential. When schools adopt blanket policies, when districts create universal rules, when we lose sight of the contextual judgment that good teaching requires, we risk reducing rich educational experiences to their most manageable components. The challenge isn’t the simplification itself—it’s remembering what gets lost in that necessary process.