Current Analysis Of Digital Accessibility In Public Service Systems, Legal Regulation, And The Necessity Of Developing Ai-Based Sign Language Virtual Interpreters

Abstract

As the digitization of public services rapidly accelerates, ensuring digital accessibility for persons with disabilities (PWDs) has emerged as a critical human rights and social inclusion challenge. This paper presents the results of an official accessibility audit of the "E-Mongolia" public service platform's website conducted in accordance with the Web Content Accessibility Guidelines (WCAG) 2.2 (Level AA) standards, and discusses the associated legal and technological challenges. The audit identified a total of 503 accessibility violations within the system, revealing a weak monitoring and enforcement framework for digital governance accessibility. Currently, the Communications Regulatory Commission of Mongolia (CRC) monitors sign language compliance only for traditional television broadcasting, while no monitoring, auditing, or enforcement mechanism exists for digital platforms, websites, and mobile applications. Furthermore, because human resources (professional sign language interpreters) are severely insufficient to guarantee the right to information for the hearing-impaired, this study argues that an Artificial Intelligence (AI)-based Mongolian Sign Language Virtual Avatar Interpreter merits urgent research and development as a conceptual, human-in-the-loop complement to human interpreters—not a replacement for them—pending empirical validation and safeguards for privacy and community participation. To compile the large-scale video dataset of sign language that such a model would require, this study proposes a national implementation model. The model involves a joint initiative by the Ministry of Digital Development, Innovation and Communications (MDDIC) and the Ministry of Labour and Social Protection (MLSP) to equip national public call centers with video-based sign language translation capabilities, leveraging crowdsourcing methods for data collection.