Hector Research Institute of Education Sciences and Psychology
AI2Teach - Artificial intelligence in foreign language teaching
AI2Teach offers individual support in foreign language teaching through an enhanced digital teaching-learning concept. With the FeedBook, an AI-supported learning environment, students receive individual support through interactive task processing with so-called scaffolded feedback, i.e. feedback that guides learners step by step to the correct solution, adapted to their level of knowledge. Teachers benefit from a dashboard that gives them a continuous overview of their students' level of knowledge.
In collaboration with the ZSL, AI2Teach offers comprehensive teacher training on the topic of artificial intelligence in English lessons. Digital tools such as the FeedBook are combined with scientific findings from educational psychology and second language acquisition research. One focus of the training is on the use of intelligent tutorial systems (ITS) in English lessons in years 7 and 8, with the web-based FeedBook serving as an exemplary ITS.
As part of AI2Teach, we are investigating how students and teachers benefit from working with the FeedBook and teacher training. We are investigating factors such as learning performance, motivation and teaching quality.
The FeedBook has been systematically developed further. Improvements should enable adaptive scaling of computer resources to make the system available to many users. In addition, improvements and enhancements to the user interface were implemented, including dashboards for teachers and students and new task interaction types. In terms of content, new exercises were designed and adaptive task selection and customization during processing was added.
In an iterative, user-oriented process, the teacher dashboard for FeedBook was developed, which gives teachers a quick overview of their students' performance and learning development at class and individual level. Functions such as a start page with statistics, an overview of the submission and correctness of compulsory tasks and a display of acquired skills and typical errors are being developed.
Conception of a techer professional development training
In collaboration with teachers, AI2Teach is developing a multi-part training course that takes into account the latest training research. The training combines key aspects such as teaching quality concepts in English lessons, subject-specific didactic principles, second language acquisition research, practice, diagnosis, individual support and digitalization, including the opportunities and limitations of current AI developments.
The newly developed system was used at secondary schools in Baden-Württemberg for several months in 2024 and will be accompanied by a randomized controlled field study. The students' learning performance was recorded using tests. Factors such as motivation, teaching quality and personality traits were recorded in questionnaires.
In addition, the students' clickstream data, which is generated when using the FeedBook, is systematically analyzed. The aim is to investigate how learners interact with the system. This will enable continuous improvement of the FeedBook.
How schools can benefit
The findings from AI2Teach will help us to better understand how teachers can be supported in the best possible way with an intelligent adaptive tutoring system and, in particular, how they can be supported in the use of AI in the classroom through comprehensive further training. Furthermore, the resultsprovidefurtherinsightsintohowstudentslearnforeignlanguages in an intelligent adaptive tutoringsystem such asFeedBook, howtheyinteractwiththesystem and whichfactorshave a positive effect on theirlearningperformance. In thefuture, thesefindings will makeit possible toadapt adaptive tutoringsystemsspecificallytotheneedsoflearners and teachers and to promote sustainableimplementation in everydayteachingthroughcomprehensive professional support.
