Reading for Understanding: An AI-Enhanced 20-Day VLI Reading Intervention for Grades 6–8

Abstract

This action research examined the implementation and preliminary outcomes of a teacher-developed, artificial intelligence (AI)-enhanced 20-day reading intervention organized around three core domains: Vocabulary, Literary Text, and Informational Text (VLI). The intervention was developed in response to identified reading needs among Grades 6–8 learners participating in a summer reading remediation program in Detroit, Michigan, United States. Instruction was delivered for approximately 70 minutes per day and incorporated explicit instruction, teacher modeling, guided and independent practice, collaborative learning activities, formative assessment, differentiation, Universal Design for Learning (UDL) principles, and AI-supported instructional material development. Program-level assessment data demonstrated positive improvement following implementation of the VLI intervention. The overall English Language Arts (ELA) mean increased from 36% on the pre-assessment to 56% on the post-assessment, representing a 20-percentage-point gain. Performance distributions likewise shifted toward higher achievement levels. The proportion of learners performing at the Below Basic level decreased from 38% to 10.8%, while the proportion reaching Proficient increased substantially from 3% to 29.2%. The Basic performance group also increased from 21% to 32.3%. These results indicate a meaningful positive shift in reading performance following the 20-day intervention. Because available assessment reports contained differing numbers of records across administrations, findings are interpreted as descriptive rather than causal. Overall, the findings provide promising preliminary evidence of the effectiveness and instructional potential of the VLI model and support its continued refinement, further evaluation, and possible contextual adaptation for inclusive reading intervention settings.