![[Translate to English:]](/storages/ifs-ep/_processed_/8/5/csm_AdobeStock_412860748_7275832d82.jpeg "[Translate to English:]")
Promoting reading skills through artificial intelligence and digital stories
LeseKI is an AI-supported programme designed to strengthen children’s reading literacy and motivation to read. On the platform lesekids.tu-dortmund.de, children can personalise stories according to their interests and preferences. The system also uses adaptive reading assessment to adjust text difficulty to each child’s abilities. The reading texts and assessment are designed for children in Grades 3 and 4.
Funding
Project description
How can artificial intelligence be used effectively to promote reading literacy? This question is at the heart of LeseKI, a project by TU Dortmund University and Goethe University Frankfurt in cooperation with the Vodafone Foundation.
Effective reading support must bring several educational goals together. It should reflect children’s interests while pursuing curriculum goals. It should respond to individual learning needs while making room for shared and cooperative learning. It should also reduce teachers’ workload while leaving pedagogical and instructional responsibility in their hands.
To address these goals, LeseKI combines adaptive reading assessment with personalised reading support. Results from short assessment tasks help the system adjust the difficulty of texts and activities to a child’s reading ability. The child’s subsequent responses, in turn, provide further information to guide support.
The programme has two connected components:
LeseKI Mission – adaptive reading assessment. Children work through short texts and tasks that assess, among other things, their understanding of difficult words and their ability to draw different types of inferences. They receive immediate feedback on their responses, including answers to open-ended questions. Task difficulty adapts to their previous responses, gradually building a picture of their reading ability.
LeseKI Themenreisen – personalised stories. Children choose from themes such as careers or adventures and help shape the characters, settings and storylines. Each story is accompanied by matching comprehension tasks. Text difficulty is informed by the adaptive assessment.
LeseKI has already been tested in a scientific study involving 40 classes. Initial findings show that, compared with non-personalised texts, children rated the personalised stories as more interesting and engaged with them more often and for longer. Whether this also leads to improvements in reading literacy remains under investigation.
A further module is planned to support children as they work through texts and tasks modelled on learning assessments and large-scale studies such as PIRLS.
The accompanying research focuses on three questions:
- How do the components of LeseKI affect reading literacy, motivation to read and reading behaviour?
- How can AI effectively reduce teachers’ workload while supporting their pedagogical decisions?
- How can individualised support and shared learning be combined in the classroom?
Publications
Ludewig, U., Dengel, A., McElvany, N., Royal, N., Rustemeier, L., Sommer, S., Spang, E., & Zierer, C. (2025, 28. Januar). Personalisierte Lesematerialien mit generativer KI in der Grundschule: Textinteresse, -schwierigkeit und Aufgabenerfolg in einer Leseförderung mit maßgeschneiderten Inhalten. Beitrag auf der GEBF 2025, Mannheim.
Royal, N., Ludewig, U., & McElvany, N. (2025). KI in der Leseförderung und sprachlichen Bildung [Organisiertes Symposium]. 20. Tagung der Fachgruppe Pädagogische Psychologie (PAEPS), Jena.
Royal, N. (2025). Individuelle Förderung durch KI: Personalisierte Texte als Mittel zur Lesekompetenzsteigerung [Vortrag im Symposium „KI in der Leseförderung und sprachlichen Bildung“]. 20. Tagung der Fachgruppe Pädagogische Psychologie (PAEPS), Jena.
Lead researcher at IFS
External project partners
- Prof. Dr. Andreas Dengel (Goethe-University Frankfurt)
- Vodafone Foundation