Beyond AI Detection: Redesigning Assessment for Secondary and Higher Education

Description

Generative AI is changing the way teachers think about assessment.

Across secondary and especially higher education, students can now produce essays, reports, and assignments that are well structured, fluent, and correctly referenced, and all of that can happen in a matter of minutes.

This does not mean that every student is cheating, but it does mean that traditional assessment tasks no longer always provide clear evidence of what a student knows, understands, or can do.

AI detection software doesn’t fix it. These kinds of tools don’t offer a reliable long-term solution. They struggle to keep up with new technologies, can produce false accusations, and may be especially problematic for students who write in a second language or express their thinking in less conventional ways.

This course introduces a practical approach to assessment redesign in the age of generative AI.

Instead of focusing on surveillance or suspicion, participants will learn how to design tasks in which AI use is either clearly limited, deliberately integrated, or made transparent through the structure of the assignment.

They will explore how to decide when AI use is appropriate, when it should be restricted, and how these expectations can be communicated clearly to students.

All of this will work with the AI Assessment Scale as a shared framework for defining levels of reasonable AI use in assessment, so that they will be able to map their own existing tasks against it.

Throughout the week, participants will redesign one real assessment from their own teaching context.

They will learn how to make tasks more authentic, situated, and applied, so that they require students to use judgement, personal reasoning, reflection, or oral explanation rather than simply submitting a polished final product.

The course will, in fact, also focus on process-based assessment, including drafts, learning logs, short oral defences, presentations, and demonstrations.

These elements will help educators make students’ thinking more visible without turning assessment into a policing exercise or creating an unsustainable marking workload.

By the end of the course, participants will be able to map their own assessments against a recognised international framework used in both schools and universities, design assessments that keep their meaning when every student has access to AI, judge where AI use is legitimate in their own subject, and replace suspicion with expectations that their students can follow.

They will leave with one of their own assessments fully rebuilt and a transparent AI-use policy to take back to their institution.

Concept by Veerle Ponnet

Learning outcomes

The course will help participants to:

• Understand why AI-detection and surveillance approaches fail, and why redesigning the task is the lasting solution;
• Recognize the limitations of AI-detection and surveillance-based approaches;
• Apply the AI Assessment Scale to state clearly where and how AI may be used in a given task;
• Redesign assessment tasks so they provide clearer evidence of students’ own learning, reasoning, and judgement;
• Design authentic, situated, and applied assessments that cannot be completed by generic AI alone;
• Build process and oral elements that make each student’s own thinking visible;
• Write marking criteria that reward reasoning and judgement over polished output;
• Reframe academic integrity around transparency and trust rather than policing;
• Communicate AI expectations to students clearly and consistently.

Tentative schedule

Day 1 – Introduction to the course

• Introduction to the course and the Academy;
• Introduction to the host city and the week’s cultural activities;
• Icebreaker activities;
• Presentations of the participants’ schools and institutions.

The New Assessment Landscape

• Why generative AI breaks traditional assessment, and why detection is not the answer;
• The shift from policing to trust: academic integrity as transparency;
• Sharing the assessment challenges participants bring from their own teaching.

Day 2 – A shared language for AI in assessment

• Introducing the AI Assessment Scale: a graduated framework from no AI to full AI use;
• Mapping your current assessments onto the scale;
• Matching the level of permitted AI to the real purpose of each task.

Day 3 – Designing authentic assessment

• What makes a task resistant to generic AI: situated, local, personal, and applied work;
• Turning a standard essay or report into a task that AI cannot complete for the student;
• Why do authentic, effortful tasks also lead to deeper and more lasting learning?.

Day 4 – Making thinking visible

• Process-based assessment: drafts, learning logs, and reflective accounts;
• Oral and applied checks: short defenses, presentations, and demonstrations;
• Verifying authorship without adding hours of marking.

Day 5 – Criteria, policy, and communication

• Writing marking criteria that credit reasoning and process;
• Drafting an AI-use policy that students can genuinely follow;
• Talking to students so the message lands as trust, not suspicion;
• Course evaluation: round-up of acquired competencies, feedback, and discussions;
• Awarding of the course Certificate of Attendance.

Day 6 – Structured cultural activities

• Structured cultural activity(ies) according to the hosting location;
• Learning journal: individual written reflection on the course experience;
• Course reflection: self-reflection and review of course materials and resources.

Download or share

Get weekly recaps on confirmed sessions!