

Teachers Strike Back Against AI Cheating
9 snips Aug 27, 2025
Jeremy Na, an innovative high school English teacher from the Bay Area, tackles the challenges of AI-assisted cheating in education. He highlights the shift from a grades-focused approach to a process-driven pedagogy, emphasizing in-class assignments and breaking essays into smaller tasks. Jeremy also discusses building trust with students and models responsible AI use, hinting at a deeper, more meaningful learning experience. With insights on the comeback of blue books for handwritten exams, he provides a refreshing perspective in the ongoing battle against modern cheating.
AI Snips
Chapters
Transcript
Episode notes
Detectors Are Signals Not Verdicts
- AI detectors work as 'smoke detectors' that flag likely AI text but are not definitive proof of cheating.
- Max Spero says tools should prompt investigation and teaching moments, not automatic punishment.
Detection Learns AI's Tell-Tale Patterns
- Detection models learn patterns by comparing human and AI writing and flag overused AI phrasing.
- Max Spero notes students must rewrite 30–40% of AI text to evade current detectors.
Set Clear AI Use Rules
- Tell students clear guardrails about acceptable AI use, like allowing brainstorming but banning full production.
- Use detectors to educate students and integrate self-checks before submission.