COI.
search Request for Consideration arrow_forward menu
close Submit Innovation
close
Education verified Verified Outcome TRL 9

AI-Secured Adaptive Testing

domain Client: Global Testing Body handshake Provider: Duolingo schedule Deploy: 2016-Present
96 Impact
Enterprise Ready
Evidence Score: 5/10
Strength: Very High

Executive Summary

ANALYST: COI RESEARCH

Disruption of the standardized testing market (TOEFL/IELTS) using an AI-proctored, computer-adaptive test that can be taken at home for $49 instead of at a center for $200+.

rate_review Analyst Verdict

"Massive accessibility unlock. By using AI for security (gaze tracking, keystroke dynamics) and adaptability (question difficulty adjusts to user), they proved high-stakes testing doesn't need physical centers. Accepted by 4,000+ institutions."

lock
Full Audit Report Available Includes Risk Register, Technical Specs & Compliance Data.

warning The Challenge

International students needed to prove English proficiency (TOEFL). This required traveling to a physical test center, paying ~$200, and waiting weeks for results. During COVID, centers closed, leaving students stranded. Security concerns prevented at-home testing previously.

psychology The Solution

Developed the Duolingo English Test (DET). It uses a webcam and microphone to record the session. AI algorithms analyze the video for cheating (looking off-screen, background voices) and the test response for pattern matching. The test is adaptive (shorter, more precise). Human proctors review AI-flagged sessions asynchronously.

settings_suggest Technical & Deployment Specs

Integrations
University Admissions Systems
Deployment Model
SaaS / Mobile
Data Classification
Biometric / Video
Estimated TCO / ROI
Low (User pays)
POC Summary ( to )

"N/A"

shield Risk Register & Mitigation

Risk Factor Severity Mitigation Strategy
Cheating Sophistication High Continuous red-teaming of AI proctor.
Bias Medium Fairness audits on diverse accents.

trending_up Impact Trajectory

Audited value realization curve

Accepted by Yale, Duke, and 4000+ others Verified Outcome
Primary KPITest duration < 1 hour (vs 3-4 hours)
Audit CycleCost reduced by ~75%

policy Compliance & Gov

  • Standards: GDPR, Testing Standards
  • Maturity (TRL): 9
  • Evidence Score: 5/10
  • Data Class: Biometric / Video

folder_shared Verified Assets

description
Verified Case Study
PDF • Version 1
lock
verified_user
Technical Audit
PDF • Audited
lock

Related Intelligence

article
AI Test Prep Tutor (Score Prediction)
View Outcome →
article
Dynamic Ad Insertion (Server-Side)
View Outcome →
article
Enterprise Conversational AI Assistant
View Outcome →
Security Architecture

The "Blind Verification" Protocol

How we verified these outcomes for Global Testing Body without exposing sensitive IP or identities.

Private
lock_person

1. Raw Evidence

Audit ID: #PRIV-594
Evidence: Direct SQL Logs
Public
public

2. Verified Asset

Outcome: Verified
Ref ID: #COI-594

Strategic Action Center

Identify your current stage and take the next step.

rocket_launch
Replicate This Success
Want similar results? Request a deployment consultation.
psychology_alt
Submit Challenge
Have a different problem? Submit your problem statement.
publish
Publish Case Study
Submit your own verified evidence.
thumb_up
Verify Impact
Audit your existing solution.