The Educator Magazine U.K. Sept-Dec 2026 issue - Magazine - Page 24
Ai is exposing a growing gap
between school exams
and university assessment
ensuring students are assessed against the
same standards.
However, traditional assessment often
focuses on the final product. While this
approach has its benefits, it can hide the
learning journey, making it difficult to
understand how students formulated their
ideas, responded to challenges, and reached
their conclusions.
Zemina Hasham,
Chief Customer Experience Officer
at Turnitin
As exam season ends and students prepare
for university, an important question is
coming into focus: are we assessing the skills
students will be expected to use next?
The transition from school to university has
always involved a shift in expectations. However, the widespread adoption of AI is making
that shift more visible than ever. Students
are moving beyond high-stakes assessment
environments that often emphasise
individual performance under controlled
conditions, into higher education settings
where learning is increasingly collaborative
and iterative
Outside the exam hall, students use AI as
part of their everyday learning. In fact, 92%
of undergraduates report using AI tools in
some form, while 41% say they use AI to
generate and refine research ideas. Whether
schools formally acknowledge it or not, AI is
becoming part of how students learn, helping
them explore ideas, organise information,
support research and refine their work.
The challenge is ensuring students are prepared to use AI responsibly and transparently
as they progress through education.
Schools and universities are moving
in different directions
School assessment continues to prioritise
consistency and comparability, particularly in
high-stakes examinations.
These approaches remain essential for
maintaining trust in outcomes and
Universities, meanwhile, are increasingly
looking beyond the final submission. As
generative AI becomes more embedded in
student workflows, a polished essay or
assignment is no longer always a reliable
indicator of learning. Instead, many institutions are placing greater emphasis on how
students develop, apply, and demonstrate
their thinking. This might involve drafts,
annotated research, or discussions that
provide greater visibility into the learning
journey.
The aim is to complement outcomes with a
better understanding of the reasoning,
judgement, and critical thinking that
underpin them.
The cost of inconsistent expectations
Students who have spent years working
within strict definitions of independent work
may arrive at university to discover a much
more nuanced set of expectations. Guidance
can vary significantly between schools,
universities, departments, and individual
educators. What is considered acceptable
in one context may be treated differently in
another.
Many institutions are still developing their
approaches. 41% of UK degree-awarding
institutions have no publicly accessible AI
policy, despite the rapid adoption of these
tools by students. In some situations, using
AI may be encouraged; in others, it may be
restricted. Students are expected to exercise
judgement, yet they do not always have the
opportunity to develop the skills universities
value, such as reflecting on decision-making
or evaluating AI outputs.
This lack of clarity creates confusion.
Research shows that one in five students
believe simply asking an AI tool for homework
tips constitutes cheating, highlighting how
unclear the boundaries of acceptable AI use
remain for many learners.
Students themselves are signalling a need
for greater clarity. Just 15% say they have
received enough guidance on how to use AI
appropriately, leaving many to navigate
complex questions around AI use without
clear direction from educators.
The conversation must move beyond a focus
on simply preventing misuse to prioritizing support for responsible, transparent AI
engagement. Greater transparency removes
uncertainty and helps students develop the
judgement they will need throughout higher
education and beyond.
Universities are already responding
Higher education is proactively addressing
these challenges by revisiting assessment
design, developing AI guidance, and
prioritizing the learning process over the final
output.
The most effective approach recognises that
AI cannot be addressed through a one-sizefits-all model. Instead, institutions should
consider a range of factors, including the
nature of the assessment, how AI is being
used within the learning process, and the
policies and learning objectives that underpin
it. What is appropriate in one context may be
ineffective in another.
Instead, institutions are increasingly focusing
on helping students understand how to use
AI effectively, where its limitations lie, and
why human judgement remains essential.
Educators play a critical role in this process.
By modelling responsible use, encouraging
reflection and creating opportunities for
students to discuss how AI has informed
their work, they can help develop the critical
thinking skills that remain at the heart
of learning regardless of the technology
involved.
Final thoughts
Schools and universities do not need identical
approaches, however, greater alignment is
needed around the skills students will require
as AI becomes an increasingly common part
of education and work.
Students need to be able to do more than
produce a strong final answer. They must be
able to explain how they reached it, evaluate
the tools they used, and demonstrate the
judgement behind their decisions.
If we want students to succeed in an AIenabled world, we must prepare them
for that reality. This requires assessment
approaches that value the completeness of
the learning process, how a student develops
their reasoning and judgement, not just the
final output.
