TeSLA — EU Horizon 2020 Trust-Based E-Assessment System for Continuous Authentication

TeSLA (Trust-based Authentication and Content Verification for E-assessment) was an EU Horizon 2020-funded project, coordinated by the Universitat Oberta de Catalunya (UOC) in Barcelona, that ran from January 2016 to March 2019 with a total budget of €7,283,092.50 (€5,916,028.50 from the EU). It brought together 18 partners — 8 universities, 3 quality-assurance agencies, 4 research centres and 3 companies — across several European countries to build a combined authentication and integrity system for online/e-learning assessment, using facial recognition, voice recognition, keystroke-dynamics analysis and forensic writing-style/plagiarism analysis, and surfacing green/orange/red "trust" alerts to instructors rather than an automatic pass/fail.

The system was piloted with 22,941 learners (including 861 students with special educational needs) and 457 teachers, across 532 assessment activities in 310 courses, through 7 large-scale institutional pilots plus 6 further third-party pilots. Per the project's own reporting, over 70% of participating students agreed the tools helped ensure their examination results were trusted. The project also produced a proposed European quality framework for online assessment intended to outlive the funding period.

TeSLA is a completed research-and-innovation project rather than an ongoing commercial deployment: EU funding ended in March 2019, and while some tooling was reportedly made available afterwards, there is no confirmed evidence of continued operation at the original pilot scale. Continuous biometric monitoring of students (face, voice, keystrokes) also carries an inherent privacy/GDPR tension that the project addressed partly through informed-consent design, but which remains a structural caution for any similar system deployed after the project's own data-protection safeguards lapse.

Read the full analysis: https://cordis.europa.eu/article/id/386857-e-assessment-system-to-eliminate-e-learning-cheating