Audemy — AI Audio-Based Adaptive Learning Platform for Blind and Visually Impaired Students

Audemy is an AI audio-based learning platform built specifically for blind and visually impaired (BVI) K-12 students, developed by then-high-school student Crystal Yang with AI-PC hardware and local on-device AI processing support from Intel.

Provide adaptively-paced, audio-first tutoring for BVI students — tuning question difficulty, content pacing and delivery format to individual need — while protecting sensitive voice data via local, on-device processing rather than the cloud.

The platform was co-designed over roughly a year through structured interviews with 20 educators specialising in accessibility, drawn from 8 schools serving BVI students; researchers coded the interview transcripts into themes that directly shaped three adaptive features: question difficulty tuned to student accuracy, content pacing tuned to how often a student needs repetition, and delivery of material in multiple formats.

Audemy now reports over 2,000 BVI student users. The only performance figure available — one student's 28% one-month math-score improvement — is a single anecdote, not a study result.

The paper's own authors note the system still cannot interpret complex emotional states, make context-sensitive decisions from audio cues, or provide tactile feedback — a well-documented co-design process for a genuinely underserved population, but with no controlled outcome study to date.