Widely read books such as The Selfish Gene are prime candidates for AI-generated essays, because plenty of summaries and analyses exist for a language model to draw on. Instructors worry, reasonably, that students may submit polished but hollow work. Detection tools remain unreliable, so the more durable response is to design assignments that reveal genuine understanding. Clear policies and thoughtful task design go further than surveillance.

Start with a transparent policy. Students should know whether AI tools are prohibited, permitted for specific tasks such as brainstorming, or allowed with disclosure. Ambiguity breeds violations, since students may assume that whatever is not forbidden is acceptable. Placing the policy in the syllabus and revisiting it at the start of each assignment reduces confusion.

Then consider what kinds of prompts are resistant to generic responses. A prompt that asks for a summary of a chapter is easy for a model to answer. One that asks students to connect a specific passage to a discussion from class, or to apply a concept to an example they encountered themselves, is harder to fake. Specificity anchors the work in the shared context of your course.

Assignment designs that surface real thinking

Process-based assignments make it easier to see a student's own work. Requiring a proposal, an annotated bibliography, an outline, and a draft creates a trail of evidence that reveals how ideas developed. Short in-class writing exercises on the same topic provide a baseline of the student's voice and understanding. Discrepancies between in-class and out-of-class work can prompt a conversation.

• Require citations of specific page numbers and short quotations from the assigned text
• Ask students to refer to a discussion, example, or lecture unique to your course
• Collect drafts and brief revision memos alongside the final essay
• Include a short oral follow-up or conference for a sample of submissions
• Use in-class writing to establish each student's baseline voice

Grading for understanding, not just polish

AI-generated text is often fluent and confident, which can mislead graders who reward surface quality. A rubric that emphasizes depth of reasoning, accuracy of specific claims, and engagement with course material helps separate genuine understanding from generic prose. For a book like this one, watch for explanations that are technically correct but shallow, or that lack the specific examples discussed in class. These can be signals worth a closer look.

When you suspect problems, approach the student with curiosity rather than accusation. Ask them to explain a paragraph aloud or to describe how they developed their argument. A student who wrote the essay can usually do so easily, while one who did not may struggle. This conversation is more informative and more fair than relying on a detection score.

Where AI can appropriately help

Not every use of AI is a violation, and some can support learning. Students might use a tool to quiz themselves on key concepts, to get feedback on the clarity of their own draft, or to brainstorm counterarguments they then evaluate. The boundary lies in whether the tool supplements the student's thinking or replaces it. Explicit examples in the policy make the boundary easier to understand.

Instructors can also use AI-assisted grading tools responsibly by aligning them to rubrics and reviewing their output before it reaches students. Being open about this use models the transparency you expect from students. It also demonstrates that technology can be applied thoughtfully rather than as a shortcut. Students tend to respond positively when expectations apply consistently.

Building a culture of integrity

Ultimately, integrity depends on whether students find the work worthwhile. Assignments that connect to real questions, provide meaningful feedback, and allow revision reduce the incentive to cut corners. Explaining why you assign the work and what skills it develops builds buy-in. Students who see the value are more likely to invest their own effort.

Invite students to help shape the policy through discussion at the start of the term. Hearing their perspectives on fairness and on the role of new tools often produces more thoughtful guidelines. Buy-in tends to increase when students feel heard. A shared understanding of the rules makes enforcement simpler and less adversarial.