Education Surveillance & The EdTech-to-Surveillance Pipeline — Research Reference

What they monitor: Webcam, audio, screen, keystroke patterns, eye tracking, room scans. ExamSoft collects 'retina or iris scan, fingerprint, voiceprint, or scan of hand or face geometry.'

2026-06-16 6 min read Research file
Contents

Proctoring Software

What they monitor: Webcam, audio, screen, keystroke patterns, eye tracking, room scans. ExamSoft collects “retina or iris scan, fingerprint, voiceprint, or scan of hand or face geometry.”

False positives: ExamSoft flagged over one-third of ~9,000 California Bar Exam online examinees (October 2020). Vast majority were not cheating. Many Lenovo users flagged en masse due to microphone access failures.

Racial bias: Black students prompted to shine more light on their faces for identity verification; white students were not.

Proctorio/Erik Johnson lawsuit: Student published code excerpts and screenshots showing how Proctorio captures room images. Proctorio’s CEO filed DMCA takedowns against Twitter, GitHub, Pastebin. EFF sued on student’s behalf. Settled March 2022 in Johnson’s favor.

EPIC complaint (December 2020): Filed against Respondus, ProctorU, Proctorio, Examity, Honorlock for unfair/deceptive practices — excessive biometric collection, opaque/biased AI.

Student protest: 60,000+ students signed anti-proctoring petitions nationwide. CUNY: 20,000+. Cal State Fullerton: 4,500+. U of Illinois dropped Proctorio. USC dropped Respondus Monitor.

School-Issued Chromebook Surveillance

GoGuardian monitors ~27 million students across 10,000+ schools. Valued at over $1 billion. Delaware and West Virginia signed statewide contracts. Monitoring continues on school devices at home, on personal WiFi, after hours. Logging into school accounts on personal computers activates monitoring.

89% of teachers report monitoring tools in use at their schools (CDT survey). 44% of teachers say students have been contacted by police as a result of monitoring. 43% say tools are used to identify discipline violations.

Gaggle

~5.8 million students monitored, ~1,500 school districts. 95% renewal rate. Scans every document, email, chat, and picture on school Google/Microsoft accounts.

LGBTQ student outing: Gaggle flagged terms “lesbian,” “gay,” “transgender” as sexual content, routing to human reviewers. Outed LGBTQ+ students to school staff and parents. A gay student who wrote privately about homophobic parents was outed to his family via Gaggle alert. Gaggle dropped LGBTQ keywords January 2023. The Trevor Project returned Gaggle’s donation.

Self-harm flagging: 142,000+ suicide/self-harm references flagged in 2020-2021; 156,000+ in 2023-2024. After-hours alerts trigger 911 welfare checks.

Post-Dobbs: Advocates raised concerns that Gaggle could flag students searching for reproductive health information including abortion access.

AI Grading and Detection Bias

Turnitin AI detection: Claims <1% document-level false positive rate. Washington Post found ~50%. University of Pittsburgh: “not yet reliable enough to be deployed without a substantial risk of false positives.” International students falsely accused based on accented English bias.

Automated essay scoring: Penalizes non-standard English, dialectal variation, culturally grounded rhetoric. ~2/3 of AES studies rely exclusively on English-language data. AI shows racial bias when grading essays — reinforces demographic disparities in academic confidence and college admissions.

Student Privacy Laws

FERPA (1974): “School official exception” allows EdTech companies to access student records without parental consent. No standard security requirements for vendors. EdTech companies not directly liable under FERPA — only schools can lose funding.

InBloom scandal (2014): Gates Foundation-sponsored student data platform for sharing data across nine states with third-party vendors without parental consent. All nine states opted out. InBloom closed. Sparked wave of state privacy laws.

California SOPIPA (2015): Prevents companies from compiling K-12 student data for marketing. 20+ states adopted vendor-focused laws modeled on it.

School-to-Prison Pipeline

Schools with SROs have arrest rates 3.5x higher than those without; up to 8x in some states (ACLU, Cops and No Counselors). Majority of school-based arrests for nonviolent offenses. 14% of EdTech companies report providing crisis responses that include contacting law enforcement at the company’s discretion. 44% of teachers say a student has been contacted by law enforcement as a result of monitoring (CDT).

Mahanoy v. B.L. (Supreme Court, 2021)

Student failed to make varsity cheerleading, posted vulgar Snapchat from off-campus on a weekend. School suspended her from JV. Supreme Court 8-1: school’s punishment of off-campus speech violated the First Amendment.

The Normalization Problem

60% of students don’t feel comfortable expressing true thoughts if monitored (CDT). Students monitored 24/7 may not ask about abusive situations, explore resources if questioning gender/sexuality, or research controversial topics.

The chilling effect is the point the book makes: Introducing surveillance in school normalizes pervasive monitoring. This generation grows up assuming total surveillance is normal — and tolerates everything in the previous chapters because they’ve never known anything else.

Disproportionate harm: Low-income students, LGBTQ+ students, disabled students, students with IEPs flagged at higher rates.

COVID as Surveillance Accelerant

Gaggle: 20% business increase. GoGuardian: expanded to 27M students. Proctoring became critical infrastructure. EPIC filed December 2020 complaint in direct response. EdTech surveillance industry grew to $3.1 billion/year with 8% projected annual growth.

Not rolled back: 89% of teachers report tools still in use after return to in-person. Delaware and West Virginia signed statewide contracts. The pandemic normalized monitoring that persisted.

Counter-Arguments

Gaggle flags 156,000+ suicide/self-harm references per year. Welfare checks reach students in crisis. Secret Service school safety analysis emphasizes threat assessment. CDC: anonymous tip lines have some evidence for preventing violence.

The evidentiary gap: No security measure — cameras, metal detectors, police, monitoring software — has been proven to reduce school shootings. Events too rare for statistical proof. Claims of “prevented” shootings are unverifiable.

Key Institutional Reports

  • ACLU “Digital Dystopia” (2023): $3.1B/year industry, efficacy claims “unsubstantiated and unverifiable.” 26% of students expressed data privacy concerns. 21% concerned about reproductive health surveillance. Disabled students “flagged simply because of the ways disabled people already exist.”
  • CDT “Off Task” (2024): 89% of teachers report monitoring in use. 9 in 10 secondary schools use student activity monitoring. Two-thirds say students got in trouble from monitoring.
  • Senators Markey & Warren (2022): First Congressional investigation of student monitoring. Called for federal action.
researchthe-ratchet
← OlderHealthcare Data Surveillance — Research Reference Newer →Digital Identity Systems — Research Reference

Get updates on the Evil Robots series

Newsletter essays on AI escape, deception, and the humans who built them.