Language and AI in Education

Language and AI in Education

Duration 12/2023 - 11/2026

In the WoLKE project, which is funded by the Baden-Württemberg Ministry of Science, Research and the Arts (MWK), AI methods are specifically analyzed in order to develop technically and didactically suitable teaching-learning formats for use in teacher training. The increasing spread of AI in education and society requires teachers to develop not only technical skills, but also a deep understanding of the ethical and didactic implications of AI tools. Without sound training, there is a risk that AI will be used without reflection, which can lead to unforeseeable consequences in the classroom. The aim of the project is to create new, curriculum-based courses that will provide future teachers with the necessary knowledge and skills for the reflective use of AI tools in language and STEM lessons. An interdisciplinary team from the fields of computational linguistics, computer science and subject didactics, working in collaboration with the Ludwigsburg University of Education (responsible for language didactics) and Schwäbisch Gmünd University of Education (responsible for STEM didactics), is developing these formats based on international research findings. The focus is on practical relevance and understanding the possibilities and limitations of AI in education. Following successful development, the courses will initially be implemented at the partner universities. Subsequently, a broader dissemination of the results is planned.

Language and AI in Education

Duration 06/2020 - 05/2026

Following a debate on the DigitalPakt Schule that focused primarily on the school infrastructure, the project poses the central question of how the fundamental possibilities of digital learning contexts can be realized in the real school context for the effective digitalization of school education. On the one hand, AI-based adaptive, interactive systems enable a genuine improvement in learning through tailored support, which makes them attractive for the often called for internal differentiation in schools. On the other hand, there has so far been little discussion about how individual digital support can be meaningfully integrated into school lessons. Teachers play a central role here, as they can obtain important diagnostic information on the basis of individual learning analytics through digital systems, but also require substantial skills for the interpretation and the resulting methodological and didactic options for action for the guiding design of the learning process. The project offers both the necessary technical expansion and systematic further training for digitally supported teaching and learning in English. The planned expansion of the school-proven intelligent tutoring system FeedBook to include a teacher interface will process the diverse information on the learning processes and individual skills of the pupils in a class in such a way that teachers can obtain the information they need to design lessons in a way that promotes learning in a short space of time. The further training of teachers supports both the concrete use of such an interface in school practice and the learning-psychological and methodological-didactic foundations required for independent interpretation. The project is being carried out in cooperation with the Center for School Quality and Teacher Training; its effectiveness will be evaluated in several stages according to the scientific state of the art.

Language and AI in Education

Duration 11/2021 - 03/2025

Reading competences in foreign languages are essential for young Europeans to gain access to information in other countries, to compare different perspectives on complex issues and to cooperate in solving problems in international contexts. The teaching of these crucial competences cannot rely on pre-packaged materials in textbooks and readers, but needs to use reading materials that (1) address current issues to build and maintain reading motivation and that (2) meet the language levels of students to develop their reading competences. In order to increase fluency in reading, it is of particular importance that the texts' linguistic features correspond to the individual learner's level of second language acquisition. Most learning environments require individualisation, and in the field of reading in a foreign language, this means that teachers need to be able to provide texts of different difficulty level to a group of students in order to effectively foster indidivual development. Therefore, the LATILL project supports foreign language teachers by developing digital tools that enable them to identify such level-adequate texts of interest for their various classes and individual learners. More specifically, the LATILL team is to create a freely accessible platform for European – and worldwide – teachers of German as a foreign and second language that provides a search and analysis function for German texts on a specific topic and CEFR level and offers supporting tools and materials for working with authentic texts. Technical advances in the field of computer linguistics allow to develop a digital tool that enables foreign language teachers to identify texts that are level-adequate in terms of morphology, syntax, lexicon, and genre. The use of such level-adequate texts in combination with a self-regulatory approach to foreign language reading and implicit lexico-grammatical learning allows for substantial quality improvement of foreign language programs.

Language and AI in Education

Duration 01/2024 - 12/2029

PostdocTEIFUN is an interdisciplinary postdoctoral college in the field of education and AI. Research is conducted into technology-supported innovations in subject-specific teaching settings and their practical application. Extended reality (XR) and artificial intelligence (AI) applications permeate almost all areas of social life. How education and teaching relate to this is the subject of intense debate. The dynamics that can currently be observed are correspondingly high: research work on the relationship between school education and the new XR and AI technologies is already available or is currently being developed. However, there is still not enough research to be able to make conclusive statements about how technology-supported innovations can best contribute to an actual and sustainable improvement in subject-specific teaching. The cooperative postdoctoral college TEIFUN, which is being run jointly by the Professional School of Education Stuttgart-Ludwigsburg (PSE) and the Tübingen School of Education (TüSE) from 2024 to 2029, is dedicated to precisely this question.

Language and AI in Education

Duration 01/2024 - 12/2029

Listening skills are known to pose a great challenge for many language learners, as their development requires a lot of exposure to spoken language, which is in many ways less accessible than written language. Authentic spoken language is often considerably above learners' comprehension level and, compared to written language, is more difficult to adapt. In addition, learners (and often teachers) have little awareness of the building blocks of successful listening, such as the ability to recognize specific phonemic contrasts or to segment continuous speech stream into words. If these abilities are not well developed, this becomes a bottleneck in accessing any top-down listening strategies, such as using contextual information or speakers' intonation, which is very important to infer speaker's intentions and attitudes. Another major issue is that successful listening comprehension depends largely on learners' vocabulary knowledge and, crucially, learners' ability to recognize the words they know in continuous speech. However, most vocabulary learning occurs in written mode, and even when listening to word pronunciation is included, words are usually presented in isolation and pronounced very clearly. Until recently, these issues have been difficult to resolve, because finding listening resources at the right level or adapting them is very time-consuming, while using authentic resources requires a lot of individual support from teachers. Honing listening skills and improving individual word recognition in speech stream is also very tedious, requires a lot of individual practice, and learners will inevitably vary in their needs and problem areas. This project focuses on harnessing AI and digital tools to support teachers and learners in listening skills development. Using our expertise in language teaching, listening skills training, and technology-assisted learning, we aim to create an intelligent adaptive tutoring system to provide plentiful and effective listening training focusing on learners' individual goals and weaknesses, which would complement classroom teaching.

Language and AI in Education

Duration 01/2025 - 05/2028

In the first project phase of ALEE, we set ourselves the challenge of designing and developing an adaptive AI-based learning platform for individual support in business lessons. We created a prototype learning platform for a selected subject area with over 700 different tasks of varying complexity and empirically evaluated it in a pilot study in the real world of education. Even after 3 years of project duration, digital support for teaching and learning in schools, universities and other educational contexts remains at the top of the political agenda, as it can concretely address the major challenges of the education system due to the substantial heterogeneity of learners and opens up new potential for teaching-learning processes at all levels of the education system (e.g. KMK 2021). Adaptive learning systems are an important building block for successful digitalization in education. Digital adaptive learning platforms enable a personalization of learning that teachers can hardly achieve in practice. They can offer individual learning paths with tasks that correspond to the individual performance level and thus also support teachers in the design of differentiated lessons. However, such systems have so far been developed primarily for mathematics and science subjects, not for the humanities or social sciences. We have begun to address this gap by researching and developing adaptivity at the interface between STEM and social sciences, specifically in economics education. Further development of the prototype of our learning platform in a second project phase towards an applicable adaptive AI-based system in economic education is therefore very promising and necessary in order to promote adaptive learning in schools on a broad scale. The innovative character of this research project lies in particular in the interdisciplinary cooperation and the special content domain, which places further demands on the AI system. The research project therefore promises a gain in knowledge that can be transferred to similar content domains and can be applied in the future far beyond the research carried out.