Inclusive Anti-Violence Digital Platform for Persons with Disabilities

Abstract

People with disabilities (PWDs) experience disproportionate levels of abuse online, but few safety tools are developed with accessible interaction and assistive-technology support. To this end, we introduce and empirically test the Inclusive Anti-Violence Digital Platform for Persons with Disabilities (IADP-PWD) as a cross-device application that integrates accessibility-first user experience with multimodal services for violence detection. Informed by a design-science approach, the system was implemented as a three-tier architecture (presentation, service, and data) complemented with an AI moderation pipeline that blends text classification, image recognition, and user report workflows. Available accessibility features, including screen-reader semantics, captioned multimedia, haptic alerts, and an Easy Read mode, were built in according to WCAG 2.2 and EN 301 549. Evaluation involved (i) an automated audit of accessibility for 50 success criteria, (ii) a usability study with 30 PWDs, and (iii) algorithmic benchmarking on a 10,000-item harassment dataset. The platform achieved 95 % adherence, with a System Usability Scale score of 86 ± 6.4, and a moderation precision/recall of 0.91/0.88, outperforming three popular safety apps on all metrics. Amongst the qualitative entries received, a greater sense of safety and control was common for customisable alert types and simplified reporting. The findings show that integrating inclusive design principles into anti-violence tech has the potential to enhance accessibility and protection for PWD communities significantly.