This article questions the conditions for creative and critical integration of generative artificial intelligence (GAI) in a context of appropriation of cultural works within the initial training of German teachers. It is based on a research-creation project carried out in 2024-2025 with students of the German MEEF master’s program, centred on the transmedia adaptation of Friedrich Dürrenmatt’s short story The Breakdown (1956). The students were asked to transform a passage of the work into an augmented narrative involving text, image and sound, in collaboration with GAI. The analysis of the students’ productions shows how co-agency between human and machine fosters not only the development of linguistic and cultural skills but also a strengthened reflective and creative posture. The students’ initial distance toward GAI gradually turned into active curiosity throughout the research-creation process, opening the way to a conscious and critical exploration of the tool.
This article examines the crafting of prompts when a Generative Artificial Intelligence (GAI) produces a master’s thesis and its implications for the teaching of academic writing. To this end, it explores the concept of “prompt literacy” through a generation experiment with ChatGPT, which leads to a proposed categorization of prompts that could inform training practices. The focus is on the distinction between result-oriented “prompt engineering” and prompts that support a broader learning process. The article concludes that “extended literacy,” which involves critical thinking and the transformation of knowledge, necessarily encompasses prompt literacy and recommends that prompting be used under educational supervision.(traduction libre)
As part of a systematic-narrative literature review (Turnbull et al., 2023), we analyzed 111 articles published in peer-reviewed journals between January 2020 and December 2024 to understand how artificial intelligence (AI) is influencing assessment in higher education. Through this process, we observed that generative AI (GenAI) can successfully replicate a human performance during certain assessment types and that GenAI misuse is sometimes challenging to detect. Thus, institutions and faculty members resort to AI-resistant examination designs and automated proctoring systems. Furthermore, GenAI is perceived as an opportunity for pedagogical support: for professors and instructors, it can assist with creating exams and grading; for students, it can provide immediate feedback and aid practice. While AI has several potential uses in assessments, many authors underscore the importance of developing training and policies.
This paper reports on the integration of a generative AI conversational assistant—the GPT Writing Coach—into an Academic and Professional Writing course. Framed within a Scholarship of Teaching and Learning approach and based on Bandura’s concept of self-efficacy, the study explores how AI may support learners’ engagement, creativity, and the development of an authentic writing voice in a foreign-language context. A mixed-methods design was employed, combining pre- and post-course administration of the SAWSES scale with a qualitative analysis of student reflections collected through a peer- and self-assessment platform. Initial findings indicate notable shifts in how students perceive feedback and the writing process, while also raising important pedagogical and ethical considerations. These observations open avenues for deeper inquiry into the role and limits of generative AI in second-language academic writing.
This study examines the creative and professional practices of e-learning designers in corporate contexts. The conceptual framework draws on both instructional design models and creativity theory, while also integrating recent research on the use of generative artificial intelligence (GenAI). Using a qualitative approach, the study is based on five semi-structured interviews conducted with designers from diverse backgrounds working in various organizations in France. The thematic content analysis of design situations highlights institutional strategies for integrating GenAI, the main stages of the design process, and a continuum of creative practices ranging from adjustments made within prescriptive frameworks to forms of expert-level collaborative creativity. The findings suggest that GenAI does not play a decisive role either in the overall process or in the dynamics of creativity. While it may support, stimulate, or even surprise, creativity remains primarily under human control and is shaped by individuals as well as organizational contexts.
This article examines the integration of generative artificial intelligence (GAI) into assessment systems in higher education, with the aim of fostering self-reflection among future educators. A case study, conducted in 2025 among three master’s programs in “Education and Training Sciences,” invited students to experiment with GAI: simulating a 30-minute interview with GAI as a research thesis advisor, then producing a written reflective analysis drawing on theory, practice, and personal experiences. A mixed analysis of the written submissions suggests that this approach promotes problematization, the alternation between theory and practice, and the incorporation of learners’ experiences. It could therefore stimulate reflexivity and professionalization.
This document presents a field-based account of an innovative critical education project on artificial intelligence (AI) conducted in three upper-elementary (cycle 3) classrooms. The aim was to move students beyond everyday uses of AI that are often magical and anthropomorphic, towards a reasoned understanding of AI as a statistical system that depends on data and human design choices (IGÉSR, 2025; UNESCO, 2021). The sequence combined structured classroom debates, unplugged activities and hands-on training of models using the Vittascience platform, drawing on the notions of algorithmic bias, black-box/white-box models and “calibrated trust” (CSEN, 2025; Oudeyer, 2024). By placing students in a position to design datasets, the project also engaged creativity in education, understood as the capacity to explore and transform learning situations with and through AI. Results show that directly experiencing supervised learning, analysing classification errors (particularly around gender stereotypes) and confronting the materiality of data help to deconstruct anthropomorphism, foster students’ creativity and establish a critical stance. Students thus move from passive consumption to an active position as data producers, trainers, and supervisors of the machine, in line with Romero’s #PPAI6 framework (Romero, 2025). We finally discuss the conditions under which such a design can be transferred (teacher training, choice of transparent tools, centrality of structured debate) in view of developing a multidisciplinary AI literacy curriculum in primary school.
