From Guesswork to Insight
How AI is helping us see dysgraphia more clearly
This week’s focus continues my deep dive into AI imaging research for dysgraphia. I had hoped to include a follow-up review of Kaligo’s phonics component, but I forgot one small logistical detail: my primary test subject, Ernest, was away on school camp for three nights. His days were filled with hut-building, raft-making and big swings. Needless to say, homework was not on the agenda.
Still, the week wasn’t lost for Navigating Dysgraphia on the home front. In place of testing, I tumbled further down the research rabbit hole, reading new scholarly work on the emerging world of AI-assisted diagnosis for dysgraphia. What I uncovered was genuinely exciting.
My starting point was Al Abadleh et al.’s (2025) literature review, “Unravelling handwriting images: deep neural models for dyslexia, dysgraphia, and other learning disabilities detections and classifications”. Like the French studies I explored in my last two posts, this review strongly endorses the role of AI in improving how we identify and classify dysgraphia—alongside the conditions that so often travel with it: dyslexia, ASD, and ADHD (Al Abadleh 2025; Bonneton-Botté et al. 2020 and 2023; Gargot 2025).
Al Abadleh et al. (2025) make an important point: current diagnostic tools clearly produce positive outcomes, but they remain “time-consuming and lack an understanding of the complex patterns associated with these conditions.” AI, by contrast, is beginning to illuminate those patterns with a nuance our existing assessments simply can’t capture.
They also point out just how tangled and overlapping these conditions often are. Children with dyslexia, for example, can find it harder to connect sounds with letters and words, which naturally affects their writing. Students with linguistic dysgraphia struggle in a different way—they know what they want to say, but turning those thoughts into written sentences is difficult. On top of that, other conditions can influence handwriting too: ADHD might lead to rushed or inconsistent writing, while autistic students may show very repetitive or rigid writing patterns.
One of the most compelling insights was the role of handwriting images: AI can analyse writing speed, erasures, pressure patterns, and letter formation far more sensitively than the human eye. This is where deep learning becomes powerful.
AI-based systems, the authors argue, can be tailored to each learner. Deep neural networks can be trained to recognise the specific handwriting “fingerprints” linked to the different types of dysgraphia—whether the difficulty lies in motor control, spelling, or turning ideas into written language. This means AI isn’t just spotting that a child has dysgraphia; it can help identify which subtype is causing the breakdown, leading to far more targeted and meaningful support. With learning disabilities presenting so differently from one child to the next, this level of personalisation isn’t just helpful—it’s essential (Al Abadleh et al., 2025).
Importantly, integrating deep learning models into educational and therapeutic settings begins to close the gap between research and everyday practice. These tools hold the potential to provide teachers, parents, and clinicians with real-time, actionable insights to guide intervention decisions (Al Abadleh et al., 2025).
From both a professional and personal standpoint, the promise of accessible, responsive diagnostic support feels like a glimpse into a future where students like Ernest no longer wait endlessly for clarity. AI won’t replace expert assessment, but it may finally illuminate the handwriting “black box” that parents, teachers, and children have struggled to understand for so long.
According to Al Abadleh et al. (2025), AI diagnostics will ultimately “allow educators, parents, and clinicians to select intervention techniques with real-time feedback and actionable information” for neurological conditions causing handwriting delays (Santhiya et al., 2023).
How exciting is that?
Let’s hope this happens soon—that the technology becomes part of everyday educational practice, delivering quicker, clearer data to guide interventions and easing the pressure on families and schools waiting for assessments.
More soon!
Alexandra
English Teacher
Ph.D., MEd.
Subscribe for free to tag along as we explore, trial and test all things evidence-based (and occasionally quirky) to support kids with dysgraphia—one hopeful step (and story) at a time. Or go paid (for less than the price of a coffee per week) and access the Navigating Dysgraphia Club: a more personalised experience with bonus tools, events, tailored guidance, and a warm community cheering you on. All proceeds go directly toward supporting and sustaining the program.
Reference
Al Abadleh, AH. Al-Shqeerat, KHA. Shaikh, MA. Wahab Sait, AR. (2025). Unravelling handwriting images: deep neural models for dyslexia, dysgraphia, and other learning disabilities detections and classifications: a literature review. PeerJ Comput. Sci. 11:e3296 DOI 10.7717/peerj-cs.3296
Bonneton-Botté, N. Fleury, S.; Girard, N. Le Magadou, M. Cherbonnier, A. Renault, M. Anquetil, E. Jamet, E. (2020). Can Tablet Apps Support the Learning of Handwriting? An Investigation of Learning Outcomes in Kindergarten Classroom. Comput. Educ. 151.
Bonneton-Botte, N. Miramand, L. Bailly, R. Pons, C. (2023). Teaching and Rehabilitation of Handwriting for Children in the Digital Age: Issues and Challenges. Children. 10(1096). https://doi.org/10.3390/children10071096.
Gargot, T. (2025). Dysgraphia. Soins. Pédiatrie, Puériculture, 46(342), 26–28. https://doi.org/10.1016/j.spp.2024.10.009