The Guardian: a rookie teacher on how chatbots are changing literature class

The Guardian has published an audio version of a personal essay by Peter Baker about how generative AI is changing the school classroom. The author came to teach English literature and quickly realized that the argument over chatbots is no longer theoretical here. It is daily, and it is very tense.

A new teacher's anxiety

Baker began teacher training at 39, after a long career as a freelance writer and novelist. He wanted to teach teenagers to read more attentively, write more precisely and engage more deeply with literature. But almost immediately he faced a question he had no ready answer to: what to do about free chatbots that can produce, on request, coherent, convincing text with a thoroughly school-like tone. For a novice teacher, this piled on top of the usual first-year worries and made AI his main professional stress.

"Adding AI to this mix was like drinking coffee during a panic attack."

He then wades into the dispute between two camps. For some teachers, AI is a cheating machine that teaches students to avoid working through difficulty, doubt and the long refinement of an idea. For others, it is a potential assistant that can give personal feedback to every student and help those stuck at home over a draft. Baker does not pick a simple ideological side. He needs not a manifesto about the future of education but a workable solution for a real literature lesson and real writing assignments.

How chatbots break the classroom

During his placement at a suburban Chicago school, he observed lessons by an experienced teacher he calls Emily. That is where the abstract debate turned into a set of daily scenes: checking suspiciously smooth essays, conversations about nonexistent quotations, and looking for the line between a student's growth and a machine polishing the text. Added pressure comes from the fact that AI is now built into almost everything: search, email, school laptops and the programs teachers themselves use.

• Fully generated schoolwork
• Quotes and references invented by AI
• Jumping from Google search straight into Gemini and AI Mode
• Chatbots that, when instructed, imitate the style of a 15-year-old student

Yet the strongest contrast Baker saw came not from technology but from its absence. When the class read All Quiet on the Western Front without laptops and phones, the text gradually came alive: students stopped complaining, began arguing about the characters' motives and actually entered the book. For the author, it was a reminder that the point of a literature lesson is not just to get an answer, but to hold attention, work through confusion together and give the text time to take effect.

Why talk about AI

Over the summer, Baker deliberately tested chatbots, including educational versions and student modes. He began with the worst case: can a model be made to write as if an ordinary teenager had produced it, with small errors and uneven style? His conclusion is unsettling: the old comfort that a teacher will always spot machine text immediately no longer holds. At the same time, some bots proved to be useful draft editors and answered clarifying questions about assignments reasonably well. So the choice no longer boils down to a naive "allow or ban."

The most productive moments in class, for him, came not from trying to build AI into every assignment but from a direct conversation about what this technology actually is. Students said they use chatbots for flashcards, test prep, clothing advice, cooking, workouts and even pet care. Many of them themselves feared that such tools erode the ability to think independently. From this Baker draws a cautious conclusion: schools need both protected zones for slow reading and writing and a separate AI literacy track, covering data, business models, the labor of moderators and the risks of the systems themselves.

What this means

The Guardian story offers no universal rule for schools, but it shows well what the first real wave of AI in education looks like. Chatbots are already too close to ignore and too powerful to let into every piece of writing without thought. The most sensible middle course right now seems to be to keep parts of learning AI-free, discuss the technology openly, and keep human reading and human response to texts at the center of the classroom.