Abstract:

Chatbots are programs that enable human-machine interaction following a conversational style (Chase, 2016). Currently, chatbots are gaining popularity because they enable adaptation and personalization in synchronous automatic communication. In education, the number of chatbots are rapidly increasing due to their potential for supporting on-demand interaction with students through dialogue.

Considering feedback and chatbot rely on dialogue-based interactions, the incorporation of chatbots in feedback processes in learning might seem a straightforward application. According to Carless and Boud (2018) and Nicol (2019), feedback is the action through which students make sense of information about their learning process, and use it for improvement. Hattie and Timperley (2007) refer to feedforward as constructive feedback happening on a continuous basis and closely linked to learning activities, leading students to implement the suggestions for improvement in further actions. Thus, feedback involves obtaining information about the quality of learning processes and products, as well as interpreting the information and designing actions for improvement (Tai, Ajjawi, Boud, Dawson and Panadero, 2018). Providing feedback through chatbots might allow students to have the support they need, helping them better understand the assessment criteria of a learning activity, or guiding them through specific indications. From this view, chatbots can foster students’ self-reflection and self-regulation skills.

To date, the main uses of chatbots in teaching and learning have focused on management, frequently asked questions, tutoring, skills practicing, simulation and assessment (Durall & Kapros, 2020; Dutta, 2017; Fryer et al., 2020; Kerlyl, Hall & Bull, 2006). To the best of our knowledge, chatbot designs that focus on feedback provision and the self-regulation of learning are scarce. Also, despite the emphasis is still in the exploration of possibilities of chatbot technologies for teaching and learning, quite often the teachers and learners - the ones who are expected to use these tools- are excluded from the design process. Participatory design approaches have already been highlighted as valuable for technology-enhanced learning since they lead to more relevant and flexible solutions, with higher adoption rates and an increased sense of ownership by the people who use the tools (Druin, 2014; Gros & Durall, 2020). Given that chatbots may transform teacher-students relations, especially in aspects concerning feedback provision, we advocate for the involvement of teachers and learners in the design process.

In this paper we present a case that uses a co-design approach to design a chatbot that supports self-regulation and feedback skills in higher education. As part of the co-design process, online workshops with teachers and students were organised in order to gain understanding of the context of use, and capture their needs, wishes and expectations regarding the integration of a chatbot in a complex learning task. Building on the analysis of the co-design outputs, we identify and define strategies for feedback and self-regulation skills that are presented in a chatbot style guide. The chatbot style guide includes key principles and guidelines for providing feedback through chatbots in order to support self-regulation skills. We consider this is a valuable contribution that can inform further research and design of chatbot tools in education contexts.

Keywords:

Competency based education, assessment of student learning, e-mentoring, educational technology, higher education.