Writing Analytics and AI for Special Education: Preliminary Results on Students with Autism Spectrum Disorder

This article discusses the utilization of writing analytics in Special Education, with a particular focus on students with Autism Spectrum Disorder (ASD). Research increasingly supports the use of data mining and Artificial Intelligence (AI) to analyze and support students’ writing processes, showcasing the potential of these systems to enhance student engagement and the accuracy of automated feedback. However, concerns persist regarding potential biases and ethical implications. The literature highlights limitations in applying Writing Analytics and AI to atypical students since most research and tools are designed with typical students in mind, reflecting societal biases. Autistic students often encounter challenges in writing performance due to factors such as rigid style, limited vocabulary, and difficulties expressing thoughts. This paper presents a study involving the analysis of 2643 essays from secondary education students, including a subset with ASD, using text-to-network tools and NLP analysis to compare texts and examine computational linguistics metrics and text mining patterns. Preliminary findings suggest the necessity for tailored evaluation and interventions for ASD students. While AI offers opportunities for personalized interventions, further research is essential to effectively adapt current tools for atypical students.