AI Proctoring Software in Schools: What Parents Need to Know About Exam Surveillance

When remote learning expanded during the pandemic, something arrived in American classrooms that most parents never noticed: software that watches your child take a test. Not a teacher. Not a proctor in a testing center. Software that scans their face every few seconds, tracks where their eyes move, analyzes their keystrokes, and flags behavioral anomalies to be reviewed later. Schools and colleges are still using these systems, and in many cases, students and parents don’t know exactly what data is being collected, how long it’s retained, or what happens when the algorithm decides a student looks suspicious. This is not a hypothetical privacy concern. It is active policy in thousands of schools right now.

Key Takeaways

• AI proctoring software like Proctorio, Respondus Monitor, and ExamSoft scans students’ faces, eye movements, room environments, and keystrokes during online exams, often requiring installation of persistent browser extensions on home computers.
• Research by Shea Swauger (2020) documented that these systems have higher false-positive rates for students with darker skin tones, and for students with disabilities whose natural behaviors (gaze patterns, movement, tics) are flagged as suspicious.
• FERPA gives students and parents important rights regarding educational records — but it is unclear whether behavioral analytics data collected by third-party proctoring vendors falls within FERPA’s protections.
• Illinois BIPA and other state biometric privacy laws may require explicit consent before collecting facial geometry data from students.
• Parents can request information about which proctoring systems their school uses, what data is collected, and how to opt out or request accommodations.

What AI Proctoring Software Actually Does

The basic concept is straightforward: if a student is taking an exam online from home, the software tries to confirm they’re not cheating. The methods AI proctoring companies use go considerably beyond that basic goal.

Face scanning and recognition. The software captures video of the student’s face throughout the exam. It verifies that the face in the frame matches the student’s ID photo. It tracks whether the student looks away from the screen, for how long, and in which direction. Looking off-screen repeatedly is flagged as suspicious.

Eye tracking. More sophisticated systems attempt to track eye movement patterns, flagging students who appear to be reading off-screen content. Eye tracking is notoriously sensitive to lighting conditions, glasses, and individual variation in how people visually process text.

Room scanning. At exam start, students are often required to do a 360-degree room scan with their webcam, showing their desk, walls, and surrounding area. The software and/or a human reviewer checks for unauthorized materials.

Keystroke and tab analysis. The software monitors whether the student switches to another browser tab or application during the exam, how their typing patterns compare to baseline, and whether there are unusual pauses.

Flagging and review. Rather than directly accusing students, most systems flag anomalies for human review. But the threshold for flagging, and the rigor of human review, varies significantly by platform and institution.

The Research Problem: Who Gets Flagged

In 2020, Shea Swauger, a researcher and librarian at the University of Colorado Denver, published “Our Bodies Encoded,” a critical analysis of AI proctoring published in the Hybrid Pedagogy journal. The paper documented a structural problem: proctoring algorithms trained predominantly on data from students in well-lit environments with lighter skin tones performed worse when students deviated from that norm.

Darker skin tones are harder for face recognition systems to detect under normal home lighting — a well-documented problem in facial recognition research (see also Buolamwini & Gebru’s Gender Shades study). When a system fails to recognize a student’s face, it may flag the session for review or, in some implementations, prevent the student from beginning the exam entirely.

Students with disabilities face additional disproportionate flagging:

• Students with ADHD may look away from the screen more frequently — a natural behavioral pattern, not evidence of cheating.
• Students with Tourette syndrome or other conditions involving involuntary movements may trigger motion detection algorithms.
• Students with anxiety disorders may show elevated heart-rate-associated behaviors that some systems monitor.
• Students who use text-to-speech software as an accessibility accommodation may be flagged for unusual tab-switching behavior.

A 2021 report from the Electronic Privacy Information Center (EPIC) filed with the FTC cited specific cases where students with documented disabilities were falsely flagged and faced academic consequences that required appeals processes to reverse.

What These Systems Require from Home Computers

This aspect of AI proctoring receives less coverage than it deserves. Installing many proctoring systems requires granting extensive permissions to browser extensions that persist after the exam ends. In some cases:

• The extension has access to the student’s camera and microphone even when the exam is not active.
• The software scans the student’s running processes to detect communication or screen-sharing applications.
• Some implementations require students to disable VPNs or specific system processes to comply with anti-cheating detection.

The Electronic Frontier Foundation (EFF) has documented specific cases where proctoring browser extensions had permissions equivalent to surveillance software, with data retention practices that were not clearly disclosed to students or families.

Legal Framework: FERPA, BIPA, and Student Rights

FERPA (Family Educational Rights and Privacy Act) gives parents of minor students — and students over 18 — the right to access and control their educational records. The complication with AI proctoring: there is significant legal ambiguity about whether behavioral analytics data (eye tracking data, keystroke logs, video recordings) collected by a third-party vendor constitutes an “educational record” under FERPA. The U.S. Department of Education has not issued definitive guidance specific to AI proctoring.

What is clear under FERPA: schools must have a legitimate educational interest in sharing student information with third-party vendors, and they must obtain agreements from those vendors about data handling. Parents can request to see all agreements between their school and proctoring vendors.

Illinois BIPA (Biometric Information Privacy Act) is currently the most protective state-level law regarding biometric data, and it explicitly covers facial geometry. BIPA requires written consent before collecting biometric identifiers and imposes strict retention and destruction requirements. Several BIPA lawsuits have been filed against proctoring companies operating in Illinois.

State-level laws: California, Texas, Washington, and New York have varying biometric data protections that may apply to school-age students. Parents in these states should specifically ask their school’s privacy officer about proctoring vendor compliance with state law.

Comparison: Major AI Proctoring Products

Product	Parent Company	Primary Methods	Documented Issues
Proctorio	Proctorio Inc.	Face detection, eye tracking, room scan, tab monitoring, keystroke analysis	EFF documented extensive permissions; student lawsuits over disability-related flagging
Respondus Monitor	Respondus Inc.	Face recognition, room scan, behavior flagging, human review	Higher failure rates reported for darker skin tones in student accounts
ExamSoft / Examplify	ExamSoft Worldwide (Turnitin)	Secure browser lockdown, face verification, flagging	Data retention practices disputed; used in bar exams and licensing exams
ProctorU	Meazure Learning	Live human proctor + AI monitoring	Hybrid model reduces false-flag rates; data breach reported 2020
Honorlock	Honorlock Inc.	AI + human review, browser extension	Reports of flagging students whose phones appeared on desk

What Parents Can Do

Ask the right questions. Contact your child’s school or district and ask:

1. Which proctoring software is being used, and for which exams?
2. What data is collected, and how long is it retained?
3. What is the school’s agreement with the vendor regarding data sharing and deletion?
4. What is the process for reporting a false-positive flag or appealing an academic integrity decision?
5. Is there an opt-out option for AI proctoring, and if so, what is the alternative?

Request disability accommodations proactively. If your child has an IEP, 504 plan, or documented disability, document in writing before any proctored exam how their disability may interact with the proctoring system. Request explicit accommodations — often, schools can arrange supervised in-person testing as an alternative.

Understand the data before consenting. If your school sends a consent form for proctoring software, read it. Look specifically for: what biometric data is collected, who has access to it, how long it is retained, and under what circumstances it can be shared (including with law enforcement).

Contact Student Privacy Compass. The Future of Privacy Forum’s Student Privacy Compass (studentprivacycompass.org) provides resources specifically for families navigating school technology privacy issues.

What to Watch For Over 3 Months

Immediately, before the next proctored exam: Check whether your child has any software installed on their home computer from a previous proctored exam. Browser extensions from proctoring companies should be removed when no longer needed — check the browser’s extension settings.

Month 1: If your child reports difficulty starting a proctored exam (face not detected, system repeatedly asking for room scans), document the issue in writing and contact the teacher and academic integrity office immediately. Do not let a false-positive incident sit unaddressed.

Month 2: Review your school’s acceptable use policy for student data. Most are publicly available on the district website. Look for how long behavioral data from remote exams is retained.

Month 3: If your state has active biometric privacy legislation, check whether any school technology compliance notices have been issued. Organizations like the ACLU and EFF track proctoring-related legal activity by state.

Red flag: a child who has been flagged by proctoring software and received an academic integrity warning without any human having reviewed the specific incident. Automated flagging should always involve human review before any academic consequence is imposed. If your school’s process skips this step, that’s worth contesting in writing.

Frequently Asked Questions

Can my child refuse to install proctoring software?

In most cases, the practical answer is that refusing means receiving a zero on the proctored exam. However, parents can request alternative assessment arrangements — particularly for students with documented disabilities — and some schools offer in-person supervised testing as an alternative. The legal right to refuse varies by state and district. Illinois BIPA requires written consent for biometric data collection, which potentially gives families more leverage.

Does AI proctoring actually prevent cheating?

The research is mixed. A 2021 study in Computers & Education found that proctoring software does reduce some forms of cheating, particularly use of notes or phones. But students motivated to cheat often find workarounds, and the systems generate substantial false positives. The deterrence effect may be achievable through less invasive means — timed exams, open-book formats, or oral assessments.

How long do proctoring companies keep my child’s video data?

It varies by vendor and contract. ExamSoft, for example, has retained exam recordings for several years after the exam in some institutional contracts. Proctorio’s data retention policies have been the subject of student lawsuits. Parents should ask their school specifically what the retention period is in the vendor agreement.

My child was flagged for cheating by AI software. What should we do?

Request the specific flagging evidence immediately — you are entitled to know what behavior triggered the flag. Request human review if it hasn’t occurred. Gather any context that explains the flagged behavior (disability, technical issues, home environment). File a written appeal with the school’s academic integrity office. If you believe the outcome is the result of biometric system bias or disability discrimination, contact the school district’s special education office or equity officer.

---

About the author

Ricky Flores is the founder of HiWave Makers and an electrical engineer with 15+ years of experience building consumer technology at Apple, Samsung, and Texas Instruments. He writes about how kids learn to build, think, and create in a tech-saturated world. Read more at hiwavemakers.com.

---

Sources

1. Swauger, S. (2020). “Our Bodies Encoded: Algorithmic Test Proctoring in Higher Education.” Hybrid Pedagogy. https://hybridpedagogy.org/our-bodies-encoded-algorithmic-test-proctoring-in-higher-education/
2. Electronic Privacy Information Center (EPIC). (2021). “Complaint and Request for Investigation of Online Proctoring Companies.” https://epic.org/documents/complaint-and-request-for-investigation-of-online-proctoring-companies/
3. Electronic Frontier Foundation. (2020). “Proctoring Tools and Student Privacy.” https://www.eff.org/deeplinks/2020/08/proctoring-tools-and-student-privacy-scrutiny
4. Future of Privacy Forum. (2022). Student Privacy Compass. https://studentprivacycompass.org
5. Buolamwini, J., & Gebru, T. (2018). “Gender Shades.” PMLR, 81. http://proceedings.mlr.press/v81/buolamwini18a.html
6. U.S. Department of Education. (2021). “FERPA and Virtual Learning.” https://studentprivacy.ed.gov
7. Illinois General Assembly. (2008). “Biometric Information Privacy Act (BIPA).” 740 ILCS 14. https://www.ilga.gov/legislation/ilcs/ilcs3.asp?ActID=3004