dislessIA — San Marino's University Trains Teachers to Use Generative AI for Dyslexia Support
San Marino
On 24 February 2024, the University of San Marino ran a ministry-accredited seminar training teachers and clinicians to use generative …
United Arab Emirates · Abu Dhabi · See the United Arab Emirates profile · See the Abu Dhabi profile
Top 59% 47/100 · Ask Evidence Copilot about this practice
A quasi-experimental study of 60 dyslexic Arabic-speaking children (8-11) in UAE public and private schools found a ChatGPT-based reading group significantly outperformed a control group on sentence-comprehension and reading-maze tests.
A quasi-experimental study by Mahmoud Gharaibeh (Department of Special Education, Al Ain University, Abu Dhabi) and Abdullah Basulayyim (Department of Special Education, King Abdulaziz University, Jeddah), published in Disability and Rehabilitation: Assistive Technology (Vol. 20, No. 8, Nov 2025, pp. 2858-2871), tested whether ChatGPT-based reading instruction could improve comprehension for Arabic-speaking children with dyslexia. The study recruited 60 children aged 8-11, diagnosed with dyslexia, from public and private schools across the UAE, and divided them into control and experimental groups.
The experimental group received ChatGPT-based reading instruction during regular school hours, while the control group followed standard instruction. Both groups sat pre- and post-tests of equivalent difficulty but different content, using the Sentence Reading Comprehension Test and the Reading Maze Test. The published results report a statistically significant improvement in reading-comprehension scores for the experimental group relative to the control group, with post-test results showing increased sentence-comprehension accuracy and better contextual understanding of reading material. The authors conclude that ChatGPT-based exercises show promise as a supplementary tool in remedial reading programmes for children with dyslexia.
The peer-reviewed abstract does not report exact percentage gains, effect sizes, or a p-value, and it does not describe how experimental and control groups were allocated (i.e., whether allocation was randomised) or how many children were in each arm, nor any explicit data-protection or informed-consent protocol for using a commercial generative-AI tool with children. These are genuine gaps that limit how strongly the results can be generalised, even though the reported direction and significance of the effect are consistent with other dyslexia-focused ChatGPT studies in Arabic.
Read the full analysis: https://europepmc.org/article/MED/40396946
Implementation detail (cost, timeline, staffing, conditions for success) is not yet available for this practice.
Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.
Where this practice's information was retrieved from, and when.
San Marino
On 24 February 2024, the University of San Marino ran a ministry-accredited seminar training teachers and clinicians to use generative …
United States of America
The first independent RCT of an AI dyslexia-intervention game found a real but modest gain for the lowest-performing US readers …
Argentina
Free open-source AAC platform for non-verbal children (autism, cerebral palsy, Down syndrome). UNICEF Innovation Fund-backed; 40+ languages; 80+ countries active. …
Sri Lanka
SLIIT researchers built Pubudu, a deep-learning mobile app screening Sri Lankan children aged 6-7 for dyslexia, dysgraphia and dyscalculia in …
Open full copilot Grounded in cited practices — always check the sources.