Co-Created and Community Owned Chatbots in PBL

The arrival of generative artificial intelligence (GenAI) technologies in education is disrupting student learning experiences, with chatbots emerging as potential learning companions leading to speculations about both new affordances and challenges (Bozkurt et al., 2023). While general-purpose chatbots such as ChatGPT and CoPilot offer dialogic affordances that can be adapted to support teaching and learning, their lack of contextual precision and transparency limits their value in particular educational contexts. In problem-based learning (PBL) environments, where collaboration and self-regulated learning are central, there is a new opportunity to explore how creative applications of chatbots can support these pedagogical tenets. Prior research suggests that applications of AI tend to mirror dominant teaching and learning approaches, which remain largely teacher-centred, and knowledge focused (Xu &amp; Ouyang, 2022). As a result, the application of AI in education has primarily focused on supporting individual and closed formats of learning rather than enhancing group dynamics and collaborative learning in more open formats, such as PBL (Otto et al., 2025; Septiani et al., 2023). In recent reviews of chatbot applications in education (Wollny et al., 2021; Hwang &amp; Chang, 2023; Labadze et al., 2023) it is found that there is a need for better alignment between learning strategies and chatbot designs. The use of instructor-led, knowledge focused GenAI applications of general-purpose chatbots in learning environments that emphasize open, contextual and collaborative learning presents several challenges. These include inaccuracies or irrelevant responses in project specific contexts derived from overly general databases (e.g. general purpose chatbots or custom chatbots based on a fix, closed curriculum), a lack of transparency both in terms of undisclosed individual use in teams, and the ‘black box’ nature of proprietary chatbots, which are neither open-source nor fully transparent about their inner workings, potentially leading to hidden biases. <br/><br/>Given these challenges and the increasing presence of AI in education, this project explores the feasibility of co-created, community-owned chatbots designed to support project teamwork. By designing chatbots with such pedagogical intent, we can learn more about the affordances and challenges of GenAI applications in learning environments grounded in sociocultural and social constructivist theories.