Why AI Detectors Are Not the Answer for Schools Grappling with Generative AI

Sarah Hanawald
Senior Director, Association for Academic Leaders, One Schoolhouse

​As Generative AI tools become ubiquitous, academic leaders are increasingly fielding questions from faculty and families about whether AI detection tools should be part of evaluating students’ work. However, before turning to these tools, it’s essential to understand their limitations and reconsider their role in the academic setting.

Understanding Faculty Concerns About Generative AI

First, it's important to recognize why faculty are concerned about the use of Generative AI in student work. The substantive writing assignments teachers in independent schools assign are designed as learning tasks, not just assessments of prior knowledge. Writing is not merely a tool for evaluation but a process intentionally designed by a teacher to deepen students' understanding and guide them to synthesize new knowledge. In other words, the process of writing, in itself, creates new knowledge for learners. Teachers (rightly) fear that the use of Generative AI shortcuts this process, cheating students out of the learning experience their teachers crafted for them. When students rely on Generative AI to complete these tasks, they bypass this crucial learning process, potentially undermining their intellectual growth.

Another concern stems from the fact that Generative AI can produce outputs that are factually incorrect, often in subtle ways that are difficult to detect. Educators worry that students, still developing their expertise, may internalize these inaccuracies, leading to misconceptions. This concern echoes the earlier skepticism toward crowd-sourced resources like Wikipedia.

Finally, ethical considerations also play a significant role in the reluctance to embrace Generative AI. The development of Large Language Models (LLMs) that power these tools is fraught with ethical issues, including intellectual property infringements and inherent biases. Some educators argue that, given these challenges, students should not interact with these technologies at all.
Given these significant challenges, why wouldn’t schools want to leverage AI detectors? As it turns out, there are three compelling reasons not to pursue detection software.

Why AI Detectors Fall Short

Given these concerns, it might seem logical to turn to AI detectors as a solution. However, there are three compelling reasons why AI detectors are not the answer. Read why here.

Previous
Previous

Empowering Your New Department Chairs from the Beginning

Next
Next

A Head’s Perspective: Exploring The Power of Generative AI in Administrative Work