黑料网911

As generative artificial intelligence (AI) prompts higher education to seek more authentic forms of assessment, reflection has become an increasingly attractive way of evidencing learners’ thinking, professional judgment and development (Peiris, 2026; Miller & Konstantinou, 2022). For example, students might complete a design project and submit a reflective account explaining the decisions they made and what they would change in future. Yet this raises an important question: As we seek to embed reflection within authentic assessment, have we inadvertently privileged reflective writing over reflective practice? This blog post argues that we must distinguish between the two and rethink how reflection is evidenced in an AI-resilient assessment landscape, moving beyond written accounts towards more authentic demonstrations of reflective thinking.

The challenges linked to authenticity are exacerbated by the rise of large language models (LLMs) – AI systems capable of understanding and generating human-like language (Cheng et al., 2024). GenAI tools can readily generate text that demonstrates many of the linguistic features associated with reflective writing, including self-analysis, critical evaluation and apparent insight. Consequently, reflective writing may no longer serve as a reliable proxy for reflective practice within assessments explicitly designed to be AI-resilient. There is a risk that learners become focused on producing the appearance of reflection rather than engaging in the underlying cognitive processes that reflection seeks to foster (see figures 1 and 2).

Figure 1: Sample reflection (written by author)

Figure 2: AI-generated reflection

If reflection is a mental process rather than a written product, then authentic assessment should allow learners to demonstrate it through a broader range of modes. Dialogic and visual reflections, as well as multimodal approaches which combine text, visuals, audio, video and artefacts, may offer richer insights into learners’ thinking while reducing reliance on polished written accounts. These approaches are not necessarily immune to GenAI, but they can shift the focus away from the production of a polished text and towards learners’ ability to explain, justify and evidence their thinking. To address this challenge, educators may need to broaden how reflection is evidenced. This blog post highlights ‘dialogic reflection’ and ‘visual reflection’.

‘If reflection is a mental process rather than a written product, then authentic assessment should allow learners to demonstrate it through a broader range of modes.’

Dialogic reflection

Dialogic reflection enables learners to demonstrate reflective thinking through discussion rather than solely through written text. For example, following a usability testing session, students might engage in a structured reflective conversation with their tutor, explaining how feedback challenged their assumptions, justifying design decisions and identifying how they would approach the next iteration differently. In an age of GenAI, such dialogue makes reflective thinking more visible, as learners must respond to questions and demonstrate ownership of their experiences. It also mirrors professional practice in fields such as UX (user experience), where reflection often takes place through retrospectives and project debriefs.

Visual reflection

Visual reflection offers another way of moving beyond the dominance of written prose. Through annotated design artefacts, visual timelines, journey maps or prototype iterations, learners can communicate how their thinking developed over time. In a UX redesign project, for instance, students might produce a visual timeline combining early sketches, wireframes, user feedback and annotated prototype revisions to show how evidence informed their decisions. While AI may assist with presentation, the value lies in the learner’s ability to explain and justify the choices represented within the visual artefact. The assessment focus, therefore, shifts from the visual product itself to the reflective reasoning it makes visible.

Final thoughts

Rethinking reflection for an AI-resilient assessment landscape is not simply about preventing inappropriate uses of GenAI. It requires educators to reconsider what is being assessed when reflection is used as evidence of learning. If reflective practice is understood as a messy process of critical thinking, sense-making and professional judgment, assessment should create opportunities for learners to demonstrate these capabilities through authentic modes beyond writing. This also requires educators to value the tacit, uncomfortable and unfinished nature of reflection, and to support learners in engaging with that process rather than simply producing polished reflective accounts.

References

Cheng, Y., Zhang, C., Zhang, Z., Meng, X., Hong, S., Li, W., Wang, Z., Wang, Z., Yin, F., Zhao, J., & He, X. (2024). Exploring large language model based intelligent agents: Definitions, methods, and prospects. Wiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery, 16(3), Article e70111.

Dewey, J. (1933). How we think: A restatement of the relation of reflective thinking to the educative process. D.C. Heath & Co Publishers.

Miller, E., & Konstantinou, I. (2022). Using reflective, authentic assessments to embed employability skills in higher education. Journal of Work-Applied Management, 14(1), 4–17.

Peiris, M. (2026). Integrating SDGs into curriculum: Enhancing the affective domain through education. Amps 2025 Prague – Education Research & Teaching Conference: Exploring Academia – From Practice to Publishing.