Duolingo English Test : Expanding access to English profociency assessment through AI-enabled testing

The Duolingo English Test (DET) is an AI-enabled language proficiency assessment designed to make English testing more accessible, flexible, and affordable for learners worldwide. Traditional English proficiency tests such as TOEFL or IELTS often require students to travel to authorized test centers, pay relatively high examination fees, and schedule test dates weeks in advance. These requirements can create barriers for students from low-income backgrounds, remote locations, or countries with limited testing infrastructure. The Duolingo English Test was developed to overcome these challenges by providing a digital, AI-powered testing system that can be taken remotely from a personal computer.

One of the core innovations of the Duolingo English Test is its adaptive testing design. Instead of presenting all test-takers with the same set of questions, the system dynamically adjusts the difficulty level of questions based on the learner’s responses. When a test-taker answers correctly, the system increases the complexity of subsequent questions; if the learner struggles, the questions become easier. This adaptive mechanism allows the platform to accurately estimate language proficiency while using fewer questions than traditional exams.

Artificial intelligence plays a critical role in evaluating test responses. The platform uses machine learning models and natural language processing techniques to analyze speaking and writing tasks. Automated scoring algorithms evaluate factors such as vocabulary usage, grammatical accuracy, coherence, and pronunciation. These AI-driven assessments are complemented by human oversight and statistical validation to ensure reliability and fairness in scoring.

Another important feature of the Duolingo English Test is its secure remote proctoring system. Because the test is conducted online, the platform uses AI-based monitoring tools to detect suspicious behavior and ensure the integrity of the examination process. The system analyzes webcam video, screen activity, and user interactions during the test. Automated proctoring algorithms flag unusual patterns such as the presence of additional individuals in the room, switching browser tabs, or inconsistent test-taking behavior. These safeguards help maintain credibility and trust in the certification process.

The accessibility of the Duolingo English Test has significantly expanded opportunities for students seeking admission to international universities. The test is recognized by thousands of educational institutions around the world and is often accepted as an alternative to traditional English proficiency exams. By allowing students to take the exam from home, the platform eliminates travel costs and reduces scheduling challenges. The test is also considerably more affordable than many traditional language certification exams, making it accessible to a wider range of learners.

The use of AI in the testing process also allows results to be delivered quickly. Test-takers typically receive their scores within a few days, enabling them to submit applications to universities or employers without long waiting periods. This rapid turnaround time further enhances the usability of the platform for students navigating tight academic application deadlines.

From a broader perspective, the Duolingo English Test demonstrates how artificial intelligence can modernize large-scale educational assessment systems. By combining adaptive testing, automated scoring, and remote proctoring technologies, the platform creates a flexible and scalable model for language certification. The approach reduces logistical barriers while maintaining rigorous evaluation standards.

Overall, the Duolingo English Test represents an important example of AI-driven innovation in educational assessment. Its accessible design, adaptive testing framework, and automated evaluation capabilities highlight how digital technologies can expand opportunities for learners worldwide while maintaining the reliability required for high-stakes language proficiency certification.
For additional context and detailed documentation of this use case, please refer to pages 23-24 in the attached Casebook.