Beyond Assessment: Designing Mathematics Curricula for the Generative AI Era
Michael Grove
Deputy Pro-Vice Chancellor, Education Policy & Academic Standards
Professor of Mathematics and Mathematics Education
The University of Birmingham, UK
As generative AI becomes increasingly embedded in higher education, mathematics educators face both unique opportunities and pressing challenges. Tools like ChatGPT and Copilot are already used by students to generate worked examples, draft solutions, and summarise concepts, often before they fully grasp the underlying reasoning. While these technologies can offer support for practice and exploration, they also risk undermining core mathematical skills if our response remains focused solely on tightening assessments or detecting misconduct.
Simply trying to ‘protect’ traditional assessments is a battle we will not win. Generative AI’s mathematical capabilities continue to improve, and detection tools are unreliable, especially in tasks involving short form reasoning or iterative problem-solving. Reliable detection of AI-generated content in mathematical contexts is unlikely ever to be possible in a consistent, equitable way. Rather than relying on detection, we must design assessments and curricula that develop and value the reasoning, understanding, and critical skills unique to human mathematical thinking. Pretending that AI is irrelevant to mathematics, or that it can be easily policed, risks ignoring both the opportunities and the realities our students face.
Instead, we need a proactive shift towards deliberate curriculum design for a generative AI-enabled world. In a newly released white paper, I propose a programme-level framework for embedding generative AI thoughtfully and ethically within mathematics education. Rather than relying on detection or prohibition, the framework encourages educators to:
- Align AI use with programme outcomes, ensuring tasks support the development of deep mathematical understanding, proof construction, and abstraction; skills that remain irreplaceable in our discipline.
- Support progression from foundational AI uses, such as creating glossaries or generating practice problems, to advanced critical engagement, including critiquing AI-generated proofs or exploring alternative strategies.
- Introduce clear, shared frameworks for ethical AI use in assessment, helping students navigate when AI can support learning and when independent reasoning is essential.
- Prioritise equity by addressing differences in students’ access to AI tools and their confidence in using them effectively.
Mathematics educators have a unique responsibility to move the conversation beyond simply trying to protect assessments. We must instead ask what we value most in a mathematics degree and how we can design learning that ensures students develop the reasoning, precision, and problem-solving skills that AI cannot replicate. Intentional curriculum design empowers students to use AI as a tool for exploration and understanding, while safeguarding the integrity and purpose of mathematical education.
I hope this framework supports ongoing discussions within the mathematical education community and invites collaboration on how best to embed generative AI in ways that enhance, rather than undermine, our discipline’s core values.
Download the full paper: Designing the Student Learning Journey: A Practical Approach to Integrating Generative AI within Higher Education paper