Core philosophy
Surveillance: use AI to monitor and flag student behavior during exams
Verification: use adaptive conversation to confirm student understanding after submission
What it requires
Chrome browser, webcam, microphone, room scan before exam, stable internet connection
Any standard browser and a short artifact-specific probe with the configured evidence mode
What it measures
Behavioral signals: eye movement, head position, audio, screen activity, secondary device usage
Comprehension: whether the student can explain, extend, and apply the ideas in their own work
False flags
Students report being flagged for normal behavior — looking away, reading questions aloud, background noise in shared housing
Session integrity stays separate from understanding evidence so instructors can review each concern type clearly
Student experience
Room scans, constant recording, AI behavioral analysis — widely described as stressful and invasive
A brief conversation about the student's own work — designed to feel constructive, not adversarial
What you get
Flagged video segments and behavioral alerts that the instructor must manually review and interpret
An evidence report showing what the student can explain and what needs instructor review