Human-driven and AI-driven mediational tool for argumentative reworking skills: an undergraduate students lab

The development of argumentative text and information comprehension (CoI) skills related to the critical reconstruction of meaning (critical thinking, CT) is important in undergraduate education. Argumentative skills are essential both personally and professionally (Colombo, 2018; Wambsganss et al., 2020; Alotto, 2021). Especially now in the era of social media and AI-mediated information (Raffaghelli & Stewart, 2020; Selwyn et al., 2021; Nguyen et al., 2022). The use of generative AI (GenAI) seems to facilitate the information fruition process (OpenAI, 2022; Lund & Wang, 2023). However, unconscious use only increases the difficulty of navigating complex information (Roose, 2022; Bozkurt et al., 2023; Lund & Wang, 2023). In this regard, argument maps (AM), commonly used to analyze analog and static texts, could be a useful tool for visualizing, understanding, and reworking multimodal and dynamic arguments and information (Alotto, 2021; Harrell, 2022). Our study used a design-based research approach with two groups to explore the role of AMs in supporting multimodal reworking. Prospective teachers were also exposed to interaction with ChatGPT, the most widely used GenAI technology, to investigate their first perception of "artificial collaboration". Parallel to the first method, we wanted to investigate the role of AI as an ally in understanding, reformulating, and rethinking argumentative perspectives. Stemming from the Vygotskian idea, mediation involves learning within the zone of proximal development (ZDP) that does not lead to simple quantitative change but generates "qualitative transformation" (Raffaghelli, 2014). Mediation occurs by proposing “double stimulation” through “cultural artifacts” that both stimulate learners’ response (or even better, activity), as well as reflection about the cultural system embedded in the artifact. In our case, the learners’ response is argumentative reworking, and the double stimulation is produced by AMs and ChatGPT as a socio-technical artifact. To better understand the type of stimulation we investigated the different responses to the mediation in terms of understanding dynamic information (CoI) and hence developing CT. Therefore we set out to understand whether and which of the two methods offered effective moments of reformulation of one's conception of processes (reworkings) and products (written productions) (Raffaghelli, 2014). Our research questions investigated three objectives: 1. 2. to investigate whether AMs support the enhancement of students' CoI and its critical reworking (CT); whether communicative interaction with the ChatGPT artificial agent supports students in reworking information (connected to CoI) and using it for critical construction of assessment tools (connected to CT). 3. Whether there are differences in the way the two tools mediate learning within the ZPD (relating to the acquisition of skills - CoI and CT). Our preliminary analysis showed that in both groups the AMs appear to have improved on average the CoI and CT proficiency levels of the elements of information after the workshop. The first look at the data collected on the interaction with the chatbot brought out a positive initial reflection of the students about the potential of ChatGPT. However, students encountered difficulties in properly collaborating with the AI. Further research could focus on constructing functional interaction prompts.