Abstract:

French is the first foreign language in primary education in Belgium. Pre-service teachers start their training with a poor language proficiency. At the same time, generative AI tools are increasingly used by students as a compensation strategy. Artificial intelligence raises a crucial question in teacher training: how can future teachers learn to use AI critically and purposefully to strengthen pupils’ learning and strengthen teachers’ self efficacy? This presentation explores well-considered choices for AI in initial teacher education.

We adopted the Design Thinking methodology, an iterative, user-centred process in five phases. Phase 1 Empathise: we analysed the needs of a pre-service teacher to achieve the required final learning outcomes and reflected upon responsible AI integration, such as bias, transparency of sources, privacy protection, copyright, environmental impact, equitable access to technology and inclusive design for learners with diverse needs. Phase 2 Define: we explored the opportunities of AI as a support for language proficiency and defined criteria for prompting to generate relevant learning materials. Phase 3 Ideate: we developed a methodology for prompting aligned to learning goals and the profile of the learner. Phase 4 Prototype: we incorporated the methodology into a learning pathway that was trialled over the course of a semester in three different student groups of 25 pre-service teachers. In phase 5 Test, students were trained to use genAI for language proficiency in French. They engaged in authentic didactic tasks in which AI served to design learning materials for their traineeship in primary schools. They reflected on how AI reshaped the roles of teacher and learner. They explored alternative strategies and tools, ensuring that digital technology remains a support rather than a dependency.

As a first result, students made more conscious choices when prompting. Thanks to a clear methodology, the learning materials were aligned to the objectives, prior knowledge and learner characteristics. It helped to critically reflect upon criteria for learning materials. As a second result, practising language skills with the support of AI felt like a safe to fail learning environment they appreciated more than feedback in the classroom. They gained confidence as they spent more time practicing. As a third result, the dilemmas learnt them to reflect and analyse before acting, which is a crucial skill for a teacher.

Quality assurance of the Design Thinking approach was ensured through critical dialogue in cross-disciplinary learning communities, involving teacher educators, researchers and AI experts in higher education, discussing on both the opportunities and limitations of AI in teacher education.

We concluded that incorporating AI in teacher practice encourages students to reflect critically on whether the generated output truly supports the intended learning goals and whether it is reliable, inclusive and appropriate for the learner group. In this way, the use of AI becomes an opportunity to strengthen analytical thinking and pedagogical decision-making. By explicitly addressing both the affordances and risks of AI, teacher education can prepare future teachers to make informed choices that enhance pupils’ learning while safeguarding academic integrity. Developing such a critical stance towards multimodal AI systems emerges as a key competence for teachers in a rapidly evolving digital society.

Keywords:

Instructional Design and Curriculum Priorities, Impact of AI on Education, Technology-Enhanced Learning.