Conversational AI for Accessible Human-Computer Interaction: A Comprehensive Survey

Abstract

Conversational AI systemsincluding dialogue agents, voice assistants, and chatbotshave emerged as a transformative interface modality for humancomputer interaction (HCI). When designed with accessibility in mind, these systems can dramatically improve digital access for users with visual, motor, cognitive, and languagerelated disabilities. This survey provides a comprehensive review of conversationalAI technologies through the lens of accessible interaction, covering three interrelated dimensions: (i) the underlying AI technologies, including large language models, speech recognition and synthesis, dialogue management, and multimodal fusion;(ii) accessibility applications across disability categories, examining how conversational interfaces address barriers faced by blind and low-vision users, users withmotor impairments, older adults, and individuals with cognitive or developmental disabilities; and (iii) design principles and evaluation methodologies for inclusive conversational systems, including universal design, participatory methods, andaccessibility-specic benchmarks. We review over 160 papers published between2017 and 2025, identify persistent challenges in personalization, error recovery, privacy, and cultural adaptation, and propose a research agenda for the next generationof accessible conversational AI.