Skip to main content

The questions we aren't hearing: Surfacing student thinking in the age of AI

PDF

Abstract

Academic help-seeking is an important self-regulated learning strategy (Adams et al., 2024; Makara & Karabenick, 2013). This poster is informed by Makara and Karabenick’s (2013) expectancy-value model of source selection and utilization, which proposes that students’ expectations about different sources and the values they attach to them interact to influence whether they seek and use support. Students typically seek help in a preferred order, beginning with independent problem-solving before seeking support from peers and, ultimately, educators (Doebling & Kazerouni, 2021; Hyrynsalmi et al., 2025; Wirtz et al., 2018).

Students are increasingly using AI as an initial source of academic help because it provides immediate, personalised support and opportunities for iterative dialogue without many of the social barriers associated with interpersonal help-seeking (Hou et al., 2024; Hyrynsalmi et al., 2025; Makara & Karabenick, 2013). AI can also respond to incomplete or poorly articulated questions and provide straightforward explanations, reducing the effort required to request and filter relevant information (Adams et al., 2024; Hou et al., 2024; Zhang & Yang, 2025). Consequently, students may trade off the quality of educator support against the accessibility and convenience of AI (Doebling & Kazerouni, 2021; Hou et al., 2024).

As students engage more frequently with immediate and low-effort sources of help, their expectations of what help-seeking should feel like may also be changing. This shift toward AI-mediated help-seeking reduces educators’ visibility of students’ questions, misconceptions, and assessment challenges, limiting opportunities to adapt teaching and provide timely support. The challenge, therefore, is in designing low-friction opportunities that make student thinking visible to educators. One such opportunity is anonymous polling, which shares multiple features that make AI attractive to students: it is immediate, low-effort, and can be completed with minimal social risk. Unlike discussion boards, emails or open-ended “muddiest point prompts,” polling allows students to indicate their level of understanding without publicly initiating help-seeking or formulating a precise question. Research suggests that anonymous polling increases participation, encourages honest responses and supports students’ metacognition, while providing educators with insight into student learning (Wood & Shirazi, 2020; Zhang et al., 2020).

This interactive poster invites discussion about how educators can respond to the reduced visibility of student thinking in an age of AI-mediated help-seeking. Anonymous polling is presented as one practical example that may be implemented across diverse higher education contexts. Participants will be invited to share practices from their own teaching contexts and consider creative, collaborative and compassionate ways to address this challenge. A range of strategies can provide insight into student thinking, but their effectiveness depends on thoughtful design and educators adapting their teaching in response to what they reveal (Wood & Shirazi, 2020). In this way, making student thinking visible may support more responsive and compassionate teaching.

 

Keywords

Academic help-seeking, AI, polling, self-regulated learning.