Ender's Game is one of the books students know they can ask a chatbot about, and many teachers are seeing generic, polished essays that read like they came from nowhere. Detection software is unreliable and can flag honest students, so it is a poor foundation for academic integrity. A better route is designing assignments that make outside generation less useful and student thinking more visible. That approach is fairer, and it improves learning.

Specificity is your best tool. Generic prompts get generic answers, whether a student or a machine writes them. A prompt that asks about a particular scene, a class discussion, or a decision the student made about which evidence to use pushes toward work that is harder to fake.

Build in process. Ask for a planning sheet, a draft with tracked changes, and a short reflection on what the student changed and why. These artifacts show how the paper came together. A student who did the thinking can explain it, and one who did not usually cannot.

In-class writing helps too. A short timed paragraph on the same topic gives you a sample of the student's own voice and skill. Compare it to the take-home essay, not to catch students out but to start a conversation if the gap is huge.

Design moves that make the work personal

Connect the essay to something only your class has experienced. Ask students to refer to a seminar moment, a peer's argument, or a lesson activity. Require a specific number of quotes with page or chapter references. These features do not eliminate outside help, but they raise the cost of skipping the thinking.

• Tie the prompt to a class discussion or an in-class activity
• Require a planning sheet and a short reflection with each essay
• Add a brief in-class writing sample to compare voice and skill
• Ask for a two-minute conference where the student explains the thesis
• Score the reasoning between quotes, not just the presence of quotes

Why detectors are a shaky foundation

AI detection tools produce false positives, and some studies have found higher error rates for writers whose first language is not English. A flag alone is not proof, and acting on one can damage a student's trust. Use any detector output, if you use it at all, as a reason to talk, not a verdict.

Set clear expectations before the assignment. Explain what kinds of AI use are allowed, if any, and what you consider off limits. Students respond better to a stated policy than to a surprise accusation.

Grading with authenticity in mind

Reward evidence of thinking. A rubric line for specific engagement with the text, such as unusual scene choices or a reading that goes beyond summary, tends to favor authentic work. Generic text often stays on the surface.

AI grading tools can also help here, if used with care. GraideMind, for instance, scores against the criteria you set, which can highlight essays that hit the structure but lack text-specific detail. That is a signal to read closely, not a conclusion.

Having the conversation

If you suspect a paper is not the student's work, ask them to walk you through it. What did they mean by this sentence? Why did they choose this scene? Most students who wrote their own essay answer easily.

Approach the talk as curiosity, not accusation. Sometimes a student used AI to fix grammar or get unstuck, and the real issue is that the expectations were unclear. Fair handling builds trust and improves the next assignment.