Digital Harf: A Clinically Integrated Multimodal AI System for Pervasive Arabic Speech and Language Therapy

Asif Azad ¹, Mohammad Sadat Hossain ¹, MD Sadik Hossain Shanto ¹, Sabri Boughorbel ¹, Abdulrhman Aljouie ¹, Bdour Alwuqaysi ¹, Yahya Bokhari ¹, Ayah Othman Sindi ², Ehsan Hoque ^1,3

¹Ministry of Defense, Riyadh, Saudi Arabia
²Ability Center, Saudi Arabia
³University of Rochester, Rochester, NY, USA

Submitted to CHI 2027 May 2026

First author

Abstract

Children with Autism Spectrum Disorder in Arabic-speaking countries face compounded barriers to effective speech and language therapy, including a shortage of specialists, limited service reach beyond urban centers, and a scarcity of culturally grounded digital therapy materials. Digital Harf is a pervasive, multimodal AI platform that extends clinician-led speech and language therapy into the home.

The platform integrates language therapy, speech intelligibility, and picture-description modules within a workflow that adapts to each child’s performance. Its Agentic Synthetic Data Engine generates culturally relevant images, prompts, and language tasks under explicit therapeutic and cultural criteria. Evaluation by 13 licensed speech-language pathologists produced a 90.1% clinical acceptance rate for generated content without manual curation, while the broader platform is undergoing randomized clinical evaluation.

Citation

Azad, A., Hossain, M. S., Shanto, M. S. H., Boughorbel, S., Aljouie, A., Alwuqaysi, B., Bokhari, Y., Sindi, A. O., & Hoque, E. (2026). Digital Harf: A Clinically Integrated Multimodal AI System for Pervasive Arabic Speech and Language Therapy. arXiv preprint arXiv:2607.27212.