Search Articles

Search Results: EdTechEthics

When AI Graded Its Own Homework: An English Professor's Classroom Experiment with ChatGPT

When AI Graded Its Own Homework: An English Professor's Classroom Experiment with ChatGPT

A University of Virginia professor conducted a semester-long experiment, allowing students to use ChatGPT for writing assignments while challenging them to decide if AI could replace human instruction. The results reveal surprising insights about AI's limitations in fostering creativity, the ethics of 'Centaur' collaboration, and why 68 out of 72 students ultimately voted to keep their professor.
The Flawed Science of AI Writing Detectors: Why the US Constitution Fails the Test

The Flawed Science of AI Writing Detectors: Why the US Constitution Fails the Test

AI writing detectors like GPTZero and OpenAI's Text Classifier frequently misidentify human-written classics such as the US Constitution as AI-generated, exposing critical flaws in their underlying technology. Experts reveal these tools rely on unreliable metrics like perplexity and burstiness, leading to high false positive rates that unfairly target students and non-native English speakers. As educators grapple with AI's role in academia, the push for detection may be a misguided solution to a deeper pedagogical challenge.