简体中文
EnglishWhat Practical Theory Says About AI and Cheating in Schools
Updated
Practical Theory (practicaltheory.org) argues that AI cheating is rising in schools and that neither of the two loudest camps in the debate—"AI will revolutionize school" and "bring back the blue book exam"—offers a workable answer. The author, writing from inside a school, calls for acknowledging how easy AI makes cheating and for finding a difficult middle ground rather than pretending lesson planning alone can solve it. This summary reflects the site's published post "AI and Showing Our Work"; it is one educator's argument, not a research finding or policy statement.
The problem as the site describes it
The post, published a little over three years after ChatGPT's launch, states that the introduction of AI into schools "isn't going so well." The author reports that every administrator they have spoken with describes a clear rise in AI cheating, and says this matches what they see at their own school, SLA. The site frames this as a source of genuine frustration rather than a hypothetical risk.
Why both camps fall short
The post divides the debate into two positions and rejects both:
| Camp | Position | Why the site says it doesn't help |
|---|---|---|
| AI optimists | AI will revolutionize school and tear down old ways of doing things | Often argued by people "not in schools every day"; ignores the threat alongside the promise |
| Blue book revivalists | Return to in-class handwritten exams so no one has to figure out whether a student wrote something | Avoids the problem rather than solving it |
The author notes that organizations and reports cited in Greg Toppo's piece are trying to find a third way, but argues the middle ground on this issue is genuinely difficult to locate.
The core argument: you can't lesson-plan your way out
The site's central claim is that expecting teachers to design lessons so engaging that no student ever wants to cheat is unrealistic. With 30 students in a class, some will like some assignments more than others, and as long as subjects are mandatory and school is compulsory, no curriculum will appeal to every kid at every moment. The post still endorses trying to build deeply interesting curricula—it just treats that as insufficient on its own.
The author adds context: students live in a world of immediate gratification and dopamine hits, while the best schools ask them to be thoughtful and do hard things. In that environment, generative AI tools function as the latest and greatest "easy button" ever created.
What the site recommends
- Own the problem. The post says plainly: "It is too easy to cheat using AI. We have to own that." Acknowledging the ease of cheating is presented as a precondition for any response.
- Don't wait for a curriculum fix. The author explicitly rejects the idea that schools can "lesson plan our way out of this."
- Keep looking for the third way. The post points to ongoing efforts to find a middle path between revolution and retreat, while being honest that the middle ground is hard to find.
How to use this
If you are a teacher, administrator, or parent trying to understand one school leader's position, the post is a short, first-person argument worth reading in full on the site. Treat it as a perspective from practice—the administrator conversations and SLA observations are the author's own account, not cited data. For a balanced view, pair it with the reports and organizations the post references, and with your own school's policies on AI use and academic integrity.