Why the Future of Assessment is a Community Endeavour

A student can now sit alone, prompt a machine and produce a polished essay, working code or a plausible solution to a complex problem in minutes. For education, that is more than an administrative headache. It challenges a familiar assumption: that a finished piece of work, submitted out of sight, is reliable evidence of what a student knows, understands or can do. The risk is not simply that students use AI to take shortcuts. It is that education becomes increasingly transactional. Knowledge becomes an optimisable commodity; assignments become outputs; teachers become evaluators; and the difficult, uncertain work through which people develop understanding is exchanged for the immediate satisfaction of a finished product.

The isolation problem

Generative AI can be useful: it can explain, prompt, translate, support accessibility and help learners explore ideas. But when it becomes the default response to difficulty, it can also displace the human settings where much learning takes place.

A student who would once have taken a half-formed question to a study group may instead ask a chatbot. A learner struggling with an argument may accept a generated answer rather than test their thinking in discussion. The feedback loop becomes fast and private but also potentially solitary.

MIT’s recent report on AI in teaching and learning warns that AI is arriving while the social foundations of campus life are already under strain. It identifies signs that AI may increase isolation, weaken students’ confidence and mastery, reduce engagement with office hours and study groups, and erode the “social contract” between students and instructors. AI-Committee-Final-Report-Aug-13.pdf

That matters because education is not merely the transfer of correct answers from an authority, human or machine, to an individual consumer. It is a shared practice through which learners develop knowledge, judgment, confidence and a capacity to work with others.

Assessment is a social act

If AI makes individual, out-of-class outputs easier to generate, the answer cannot simply be an arms race of AI detectors, lockdown browsers and adversarial policing.

Those approaches may have a limited role in specific high-stakes contexts. But they do not solve the more fundamental problem: how to design assessment that helps learners do the intellectual work for themselves, while giving educators credible evidence of learning.

MIT cautions against relying on AI detectors, noting both their limitations and the risk that policing creates distrust between educators and students. Its alternative is not to retreat from assessment, but to make it more intentional: clearer expectations, more process evidence, appropriate in-person activity, portfolios, conversations and project-based work.

This is where assessment communities matter.

Standards do not reside only in a rubric, a grade descriptor or a marking scheme. They are created and sustained when people look at work together, compare it, discuss what makes it strong, challenge one another’s interpretations and develop a shared understanding of quality.

From individual output to shared judgment

Imagine moving beyond a model in which a student works alone and submits a single final output to an anonymous marker.

Instead, learners, teachers and moderators could work with shared examples of authentic work; compare different approaches; discuss what distinguishes a stronger response from a weaker one; and make the criteria for quality more visible.

Comparative judgment can give this work a rigorous structure. It does not mean that students must mark one another, nor that every assessment should be based on comparison. It means recognising that judgments about complex work become more reliable and standards more intelligible when they are formed through structured comparison.

When learners are meaningfully involved in recognising quality, three things can happen:

• Quality becomes visible. Learners see that good work is more than a checklist: it involves choices, trade-offs, disciplinary understanding and purpose.
• Assessment becomes formative. Comparing work can help learners internalise standards, rather than simply receiving a grade after the fact.
• Judgment becomes a capability. In a world full of polished, synthetic content, the capacity to evaluate claims, evidence, alternatives and consequences is at least as important as the capacity to generate information.

What a community approach looks like

This need not be abstract.

A department redesigning an AI-enabled design project, for example, might curate a small set of anonymised portfolios. Each could include the final artefact, a short rationale, a record of important decisions and a declaration of how AI was used.

Staff could compare the portfolios and discuss what they value: originality, technical execution, use of evidence, critical evaluation of AI output, practical effectiveness or clarity of reasoning. Students could engage with selected exemplars to understand what quality looks like in context. The team could then use what it learns to refine the task, the guidance and the assessment approach for the next cohort.

The result is more than a grade. It is a living, shared standard.

Technology should strengthen the community

The purpose of assessment technology should not be to replace professional judgment, or to create a more sophisticated system for surveilling students. It should be to make community judgment more practical at scale.

That means helping educators and learners to:

• Share and discuss exemplars of authentic work.
• Make expectations and standards visible.
• Compare judgments and surface disagreement constructively.
• Build confidence through calibration and moderation.
• Learn from assessment evidence and improve the next version of the task.

This does not remove the need for clear AI policies, thoughtful task design or individual accountability. Nor does it make every assessment “AI-proof.” It does something more useful: it makes assessment less dependent on a single polished output and more connected to the human processes through which understanding and judgment develop.

The future is shared

The future of assessment is not about finding smarter ways to police what students do alone in the dark.

It is about bringing learning back into view: through shared standards, authentic dialogue, transparent expectations and communities that can recognise quality together.

In an AI-enabled world, the most valuable outcome of education is not merely a credential or a piece of generated work. It is the development of human judgment formed, tested and strengthened in community.