Signal Behaviorally Adaptive Visual Diversion for Inclusive and Resilient Digital Assessment Delivery

Summary

Gupta Lovi Raj, Kaur Kamalpreet, Dama Sriram, and Parali Prajithaa published a paper on arXiv (cs.CY) on August 4, 2026, introducing Behaviorally-Adaptive Visual Diversion (BAVD), a theoretical framework addressing accessibility gaps in digital assessment security. The authors note that browser lockdown, webcam monitoring, and behavioral analytics used to secure high-stakes digital assessments are commonly designed and evaluated independently and often overlook learner accessibility. BAVD composites a synthetic, non-semantic visual field with assessment content, adaptively modulated based on observed candidate behavior, without ever altering the underlying assessment content itself, aiming to reduce the usefulness of unauthorized screen capture or sharing while remaining minimally intrusive for legitimate candidates. The framework includes an accessibility-aware attenuation mechanism that reduces or suppresses diversion intensity for candidates with approved visual-processing accommodations. The authors formulate the model using a coupled dynamical-systems representation and establish theoretical properties for content fidelity, rendering stability, and adaptation stability, explicitly stating the threat model and limitations, but leave empirical validation for future work.