Reading about assessment in the age of AI and a bit of Biesta

Most assessment practices are really not up to the job, but instead of completely rethinking, most often existing assessments are only tweaked and only marginally improved. Thomas & Widdop Quinton (2026) propose a way to do things differently by co-creating assessment in a Design Thinking-informed process in teacher education.

They go through five steps:

• empathizing: understanding the end-users needs and concerns. In this case in relation to assessment, through interviews with teacher educators, pre-service teachers, and high-school teachers, and thematic analysis of the data, compiled into a persona.
• defining problem: clarifying and prioritizing, leading to a problem statement: “To produce an innovative assessment framework that is engaging and provides a deeper approach to learning but does not increase the amount of time it takes to mark“.
• ideating: using “How might we…” and other brainstorming techniques to come up with radical design alternatives, and afterwards ranking them to determine the best ones, in this case with some of the teacher educators who had previously been interviewed.
• prototyping: building a prototype, and then interviewing end-users to  find out what they liked, what issues they foresaw, and to suggest changes, iteratively tweaking the prototype throughout the process.
• testing: on pre-service teachers.

The testing on groups of students who had and had not done “normal” assessments before shows that the two are perceived similar in many regards, but that the new one seems more difficult. The authors discuss that “the value of [Design Thinking] within this study was not that it produced a perfected assessment solution, but that it created a process through which assumptions could be surfaced, challenged, and iteratively refined“.

They explain that “[pre-service teacher]s identified frustrations relating to assessment clarity, workload, pacing, and perceived complexity, while academics and teachers simultaneously sought to strengthen authenticity, cognitive challenge, and professional relevance within the assessment task” and those priorities don’t always align. The assessment might have become more meaningful and authentic, but at the same time more difficult (which I think is pretty common: If you make something more interesting, it tends to become more challenging, too).

It’s interesting that the co-creation process changed the function of assessment: “assessment is not viewed solely as a mechanism for judging performance at the conclusion of learning, but as a pedagogical practice embedded within learning itself“, and I think that is something to build on more! Also the way they worked with empathy and created a persona for whom they could then design the assessment is a really useful idea to keep in mind. We’ve worked with personas developing our MOOC, but now I am keen to also try it in other contexts!

A second assessment paper I’ve had on my to-read pile is Elkington (2026), who also stresses that “meaningful assessment must continue to prioritise the human interpretive, relational, and ethical work that underpins learning“, rather than intensified surveillance, more restrictive regulations, or detection attempts in response to AI (which is part of our reality now, whether we like it or not). This can create a conflict for teachers where what is happening in constructive and trusting personal relationships with students based on their professional judgment, and what the institution wants in terms of control are misaligned. And I think it’s really important to be aware of this: “For educators, then, the challenge is not merely technical – that is, how to incorporate or regulate AI. It is also organisational and relational: how to negotiate shared concerns and develop coherent approaches to assessment […]. The result is a working environment in which forming collective pedagogical purpose – articulating what matters most in assessment – becomes increasingly challenging“. Therefore, “the response to AI cannot be reduced to policy adjustments or technical fixes. Instead, it requires deliberate attention to the social and relational work of building shared understanding and purpose within educational communities“. Approaches need to be pragmatic, using professional judgment in responding to local context of how students use AI and what that does to their learning and development as learners.

Elkington (2026) then describes a shift: a “movement from educators as custodians of knowledge, to educators and students collaborating with digital systems to augment human capacity“. This includes a shifting focus from endproduct to the process in and the considerations under which thinking happens: “Assessment, therefore, cannot merely evaluate end products or efficiently generated outputs; it must create space for students to practise and evidence precisely those aspects of thinking and meaning making that cannot (or should not) be automated“. And we need to assess how students are working with AI in meaning making: “assessments must evolve to recognise the capabilities that emerge when learners work with AI tools: dialogic engagement, iterative refinement, critical comparison, ethical reasoning and the capacity to integrate machine–generated material into coherent, situated arguments. Rather than policing tool use, human–centred assessment needs to focus on evaluating the quality of judgement demonstrated in navigating these complex interactions“. And for this, students need to learn some type of AI literacy: “This requires assessments that help students learn not only how to use AI tools, but how to discern when, why and for what purposes to use them. […] students must learn to critique AI outputs, compare them against disciplinary standards and calibrate their own judgement; equally, they must grapple with broader societal questions about data provenance, bias and equity, all of which shape the ethical landscape of working with generative technologies“.

Next, Elkington (2026) writes “assessment is ‘subject-forming’ (Biesta 2013): it involves attentional work (what students learn to notice), interpretive work (how they make sense of what they notice), ethical work (how they decide what matters), and relational work (how they position themselves in relation to others and to the world)“, and that sounds like I should read Biesta 2013 in the original, so short tangent here…

In “Responsive or responsible? Democratic education for the global networked society“, Biesta (2013) distinguishes between the responsive response (“where education is urged to adapt itself to the demands of the global
networked society“, for example by focussing on skills needed for future employment without questioning whether that is the right goal or whether a different kind of society should be created) to the world getting more and more networked, and the responsible one (which is the one he recommends because, “to a certain extent, the school also needs to be dysfunctional in relation to society because, as the French educationalist Philip Meirieu (2008) has argued, education always also comes with a ‘duty to resist’ (‘un devoir de résister’).“). He describes that the current capitalist system in a global networked society “has a tendency to create and perpetuate inequalities, and […] may result in networks that are extremely vulnerable and volatile, which raises the question not only of how much we should ‘invest’ in such networks (and how much we should invest ourselves in such networks), but also of whether there are more sustainable alternatives.”

Biesta (2013) then explains the three purposes of education: qualification, socialisation, and lastly, subjectification, i.e. person formation towards becoming “subject of action and responsibility“. And a responsible response is, according to Biesta (2013), explicitly political, and it does start from the question of how one can prepare students for the realities of the world — while, importantly, not accepting those realities as facts that cannot be changed. That is not going to be easy: “the old saying that if you stand for nothing you will fall for anything is educationally very relevant, because to resist the temptations of the global networked society – or at least to make engagement with aspects of the global networked society the outcome of a deliberate decision rather than just an automatic reflex – requires indeed what we might refer to as a certain ‘strengthening’ of the subject” because it requires “to see education as assisting the process where we rise ‘above’ our desires by always exploring whether what I do desire is also what I should desire – that is, whether what is de facto desired is also desirable“. Strengthening the subject means resisting adaptation to what is (which can go too far if not balanced by efforts to engage with the world), and Biesta (2013) states that “we need a school that is closed towards society but open towards the world. It is a school that is shielded off from direct demands from society so that an engagement with the world as a world of possibilities, a world of alternatives, becomes possible. It is the school as a space of deferment and suspension, a moratorium.”

Connecting these thoughts to democracy, Biesta (2013) explains that “a democratic orientation towards freedom is not simply about maximising one’s own freedom but is about maximising the freedom of everyone – which means that in a sense the democratic orientation is first and foremost an orientation towards the freedom of others“, and “both education and democracy come with a requirement not simply to accept what is desired or preferred, but to transform such desires and preferences into what legitimately can be seen as desirable” (my emphasis, love this quote!).

Anyway, back to Elkington (2026) and assessment: They write that “students require some agency in negotiating assessment modes, formats, contexts, and pathways in order for authenticity work to unfold meaningfully” and that includes a broad approach to assessment: “Assemblage thinking prompts us to ask not only ‘what is the assessment task?’, but also, and more instructively, ‘how does the configuration of tools, spaces, norms, and relations invite or inhibit the forms of attentional, interpretive, ethical, and relational work that assessment must support“, where AI is only element among many “whose influence must be understood, oriented, and made pedagogically productive.”

We then get a lot of practical suggestions for how to think so that “these design choices shape new regimes of engagement wherein AI becomes neither threat nor solution but one actor within the dynamic constellation of assessment practice“, for example about the shift from task design to practice design (What spaces, tools, platforms, artifacts, people shape the students’ interaction with the assessment, and how could all of them be changed to strengthen learning? How can the workspace layout, collaboration tools, resources be changed to better support students in developing the skills we want them to practice? How can we make sure that access, participation, and demonstration of learning are equitable?), participatory multimodal assessment (Does oral, written, digital, multimodal, in person, hybrid, … support learning the most? Is learning best supported by feedback in audio, video, written, annotations, automated, conversational form? (And how can I deliver it sustainably?) What constraints like accessibility, digital literacy, schedules, spaces, policies, … need to be addressed so students can fully participate?), and scaffolding (What constraints would help guide students towards deeper learning rather than just restricting them? What tools that would support learning could be integrated and how? What norms, ethical considerations, time pressures, roles, data sets can be used to make the assessment more authentic while still achievable?).

In this, I really enjoyed the discussion of artifacts (“multimodal artefacts do not merely express learning; they stage the transitions between modes of engagement that constitute meaningful educational judgement. They serve as ‘mediating devices’ that stabilise shared purposes and help students understand not only what is valued but why, enabling them to locate their own creative investments within a broader landscape of academic legitimacy“) and the importance of scaffolding (“constraints stabilise regimes of planning by providing structured pathways […] They support regimes of exploration by creating low stakes spaces […] They enable familiarity by establishing predictable routines and relational encounters […] that anchor students’ sense of belonging and safety as they negotiate unfamiliar media. And they support regimes of justification by making explicit the criteria, rationale, and public standards through which students’ multimodal decisions must ultimately be defended.“)

Anyway — as interesting as that was, that is enough co-created assessment practices for today!

Biesta, G. (2013). Responsive or responsible? Democratic education for the global networked society. Policy Futures in Education, 11(6), 733-744.

Elkington, S. (2026). Assessing what matters most – the shifting value proposition of human-centred work in assessment practice. Teaching in Higher Education, 1–18. https://doi.org/10.1080/13562517.2026.2680077

Thomas, M. B., & Widdop Quinton, H. (2026). Reimagining assessment design in higher education through Design Thinking. European Journal of Teacher Education, 1–24. https://doi.org/10.1080/02619768.2026.2705359

Pictures from a beautiful walk & talk with Rachel and Kirsty two weeks ago.

A little geology never hurts, like here a fossilised mud volcano!

But also just a beautiful beautiful coast.

And eel nets, as I leaned.

Trees really coast to the sea are also always awesome!

Especially sunny trees

Safety first!

It feels like such cognitive dissonance to have so many trees so close to the sea.

A little harbour!