While writing has been the dominant mode in academia for a long time, multimodal social semiotics (e.g. Kress, 2003; 2010; Huang & Archer, 2017) and designs for learning (e.g. Selander & Kress, 2021) provide arguments for approaching other modes as viable resources for meaning-making and knowledge production, even in higher education. This presentation introduces a recent pilot study regarding possibilities and challenges in students’ production of multimodal degree theses within the teacher education in visual arts. The project was motivated by an interest in developing approaches to knowledge production and knowledge representation that acknowledge, and take advantage of, the visual as a valid mode and of epistemological qualities in transductive and transformative processes, while also providing possibilities for inclusion in a wide sense. The data consists of five multimodal and digitally disseminated degree theses (using Research Catalogue as a platform) and conversations with the five students who produced them. The analytical framework stem from multimodal social semiotics and designs for learning, analyzing students’ multimodal knowledge production and -representation (conceptualized as designs in learning) and the educational contexts framing their work (i.e. designs for learning). Preliminary findings suggest a number of positive effects, e.g. in relation to academic literacy, understanding of subject matter, social inclusion and identity; while also implying some challenges in terms of workload and construing cohesion in the multimodal text. In conclusion, the approach applied in this project appears as a possible first step towards a wider recognition of multimodal forms of knowledge representation in academia, while also indicating some issues that need further consideration.