AI and assessment: the challenges and opportunities for international educators

Almost every process in education is being impacted by the rise of AI, and assessment is no different.

Now that students are able to use AI tools to write essays, solve complex maths problems, and create presentations, the work that they submit tells teachers less than it once did about the skills and knowledge the student has developed during a project. It is harder to determine what a student actually knows - and where gaps still exist in their understanding.

As educators, this presents major challenges, but also opportunities. In the age of AI, how can we accurately assess students’ skills? How do we evaluate the creativity, critical thinking and understanding behind a piece of work?

If your school is grappling with the challenges of AI and assessment, this guide is the first of three posts on assessing real learning when AI is available to every student. Future articles will cover AI and formative assessment, as well as how to design tasks that measure students’ understanding rather than output. You can also read our published guide on how to develop policies for AI in education.

AI is changing the way we assess student understanding

Assessment remains a fundamental part of the learning process, allowing us to gauge student growth and identify gaps that need to be addressed. It also shapes our teaching by acting as a live feedback system that reveals what students have learned, where they are struggling and what gaps in understanding teachers need to tackle.

However, now that students can prompt LLMs to create essays and answer problems set as homework, assessing submitted work no longer gives teachers a gauge of effort, progress, and understanding.

For example, a student can use generative AI to draft a detailed history report or analyse a complex literary text in seconds, complete with structured arguments and citations. In areas like graphic design, data analysis and language translation, students can use automated tools to generate finished assets and solutions that closely mirror authentic human skill. This unchecked use of AI leaves us with the challenge of distinguishing human effort and knowledge from machine-generated output.

Some educators might rely on AI detection software to verify student work. However, research indicates these tools are unreliable, frequently misidentifying genuine human writing as AI-generated while missing sophisticated synthetic text. They also exhibit bias against non-native English writers. Attempting to enforce traditional assessment boundaries through detection seems an unwinnable battle, especially as AI content generation rapidly advances. Technical detection may soon be entirely ineffective.

As more students use AI in education and detecting AI content becomes increasingly difficult, how can we assess what our students actually know?

A layer beneath formal assessment: rethinking AI and assessment

The answer lies in establishing a continuous, in-class layer of informal assessment that functions as a real-time loop for checking understanding during lessons, rather than after class. This layer sits underneath formal assessment, and teachers aren’t checking outputs like essays or presentations, instead observing how students interact with sources, approach problems and discuss their ideas with other students.

Moving the focus of assessment from end products to the process of growth and learning allows us to see how students apply feedback and how their ideas evolve, as well as giving an insight into areas where they might need more support.

This method of evaluating student growth also mirrors real-life problem-solving, which relies on ongoing iteration, feedback, research and collaboration rather than a single final performance or output. In a world where AI can generate answers in seconds, helping our students develop the ability to evaluate and critique information becomes more important than ever. These processes and skills are exactly what AI tools struggle to replicate.

We might even consider assessing their ability to use AI tools effectively and ethically as part of the learning process. This will help to prepare students for the future workplace, where people will need to work alongside AI tools. Certainly, there are many ways we can use AI to fuel students’ imagination.

Shifting our approach to assessment will require teachers to use different strategies to assess learning. In this AI and assessment series, we’ll go on to examine two core strategies in depth: responsive formative assessment that captures student thinking in the moment, and intentional task redesign focused on evaluating authentic comprehension over simple output.

AI and assessment: what it means for teachers

The rise of AI brings long-standing questions about conventional evaluation methods to the forefront, and requires teachers to move beyond approaches that AI can easily replicate. Instead, it’s time to develop assessment strategies which are focused on fostering and evaluating deeper understanding for students.

And for teachers, it’s an opportunity to enhance your evaluation of students and, instead of focusing on their output, helping them to develop the skills they truly need from their education, which will help them thrive beyond school: lifelong learning skills, critical thinking and the courage to be creative.

In our upcoming posts, we will dive into practical, manageable strategies to help you navigate this transition with confidence. Ultimately, adapting assessment to the era of AI is not about dismantling rigorous evaluation, but refocusing it on genuine student learning. By measuring the learning process alongside student ouput, we can ensure that assessment remains a meaningful tool for growth and true understanding.

As Timothy Dillon, our School Innovator of the Year 2026, reminds us, we should embrace the arrival of AI and leverage it to amplify human connection. By combining continuous classroom check-ins with thoughtful task design, we can empower students to develop the critical thinking, creativity and independent skills they need to thrive in an AI-powered world.