The Velocity Gap - AI and the school mathematics curriculum.
AI and the school mathematics curriculum: Part 1 of 4.
AI, Curriculum and School Mathematics: A Four-Part Series
Artificial intelligence (AI) is becoming increasingly part of the contemporary landscape of curriculum, pedagogy and assessment, with mathematics education very much in the frame.
Recent AI models (2025–26) have markedly greater mathematical capacity and explanatory capability than earlier systems as they can combine explanation, multi-step reasoning, symbolic and numeric tool use, code execution, and iterative checking within a single workflow.
This raises some important questions:
What should students know and be able to do themselves?
What roles could AI play in mathematics classrooms?
How could curriculum, pedagogy, and assessment respond?
How might teachers be supported to work effectively in this context?
What mathematical knowledge, skills, processes and proficiencies are required in an AI-enabled world?
This four-part series explores some key aspects of these questions.
Part 1: The Velocity Gap
How current can a curriculum remain when the world it serves is changing faster than the review processes designed to maintain it?
Part 2: Mathematical agency in an age of AI
If AI is now part of the mathematical landscape, what does it mean to learn and do mathematics at school?
Part 3: The evidence question - AI and assessment in mathematics
What should count as evidence of mathematical learning when AI can now generate substantial parts of mathematical work?
Part 4: The teaching challenge - AI and classroom practice
How can schools use AI in ways that strengthen mathematical learning without outsourcing the thinking?
Taken together, these four pieces argue that the central issue is not whether school mathematics education can simply ‘keep up’ with AI. It is whether it can remain thoughtful, coherent, and pragmatic while the conditions around it are changing.
The Velocity Gap - AI and the school mathematics curriculum: Part 1 of 4
How current can a curriculum remain when the world it serves is changing faster than the review processes designed to maintain it?
Under current school structures, a child starting school in 2026 would likely complete their final years of secondary education in 2038–39, some 12–13 years later.
Given the broader societal impacts of AI, including changes to work, information and trust, and the pace of ongoing development, the conditions in which students learn and use mathematics are likely to change substantially over that period.
This is where the mathematics curriculum, with its 5–10 year cycle of development, implementation, evaluation and review, faces a challenge from AI.
In mathematics curriculum design, coherence, consistency, and sequencing are important.
Curriculum cycles go for several years for good reason: student learning requires progression and continuity within and across the stages of schooling, and schools and teachers need time to build experience and expertise with the curriculum and its implementation.
The pace of AI development has introduced a particular kind of velocity gap: significant changes in capability, access, and use are arriving faster than traditional curriculum review cycles are designed to handle.
The issue is not simply that students now have access to AI tools, but that the rate and scope of change will likely affect not only how teachers teach, and students learn and do mathematics, but also the rationale and aims for school mathematics itself.
AI capability may now be changing on a scale of months, while curriculum review typically operates on a scale of years. That difference is the velocity gap.
Curriculum is much more than content
Curriculum is often principally thought about in terms of content to be covered, and the sequencing and progression of this coverage.
Yet curriculum does more than specify content. It reflects choices about what mathematics matters, what students should know and be able to do themselves, what role tools such as calculators, software and AI should play, and what kinds of thinking and understanding are prioritised. These choices are not merely technical decisions; they express beliefs and values about the nature and purpose of mathematics education.
For that reason, AI has implications beyond the practical question of classroom use. It brings some long-standing questions into sharper focus:
What mathematics should students learn to do for themselves?
How should the balance between procedural competence, conceptual understanding, and processes such as modelling and reasoning be developed?
What counts as meaningful work when increasingly mathematically capable AI systems can generate answers, methods and explanations so readily?
These questions reach fundamentally into the realm of curriculum design across the years of schooling.
As AI begins to affect not only how mathematical work is carried out, but also how problems are framed, evidence is interpreted, and decisions are justified, the question is not only how curriculum responds to faster change, but how it determines what should remain central when the conditions of mathematical work are themselves shifting.
With AI tools increasingly handling procedural and explanatory work, interpretation, verification, then judgement and awareness of model limitations become even more prominent within what mathematical capability looks like in practice. Curriculum will need to operate with less certainty about the future uses and purposes of mathematics than has previously been the case, while incorporating mechanisms for agility and responsiveness in relation to developments in AI technology and their societal impacts.
Across the stages of schooling
The curriculum implications of AI are not the same across primary, secondary, and senior secondary schooling.
Foundational knowledge remains central in the primary years; abstraction, interpretation, and method choice become more prominent in the middle years; and modelling, justification, verification, and judgement take on greater importance in senior secondary mathematics.
Any response therefore needs to preserve progression and coherence across the stages of schooling.
Why previous approaches may no longer suffice
In some respects, the relationship between new technologies and school mathematics is not an entirely new issue.
Over the period from about 1970 to 2020, a range of computational technologies were progressively incorporated into school mathematics education. Calculators became routine in arithmetic contexts, scientific calculators became common in secondary school, and the use of graphing and CAS calculators and other tools, such as spreadsheets and geometry software, became part of senior secondary mathematics.
These technologies did not remove the need for mathematical knowledge. Rather, they shifted the balance of what students did mentally or by hand and what they did with tools.
They are powerful tools, and their capabilities are well understood and can be planned for over time in terms of mathematics curriculum implementation. Importantly, students require sound mathematical knowledge and understanding to pose questions, formulate problems, specify computations, and interpret results.
The present AI wave is different not only in scale but also in character. The latest AI systems of 2025–26 have explanatory capability, recognise patterns, apply reasoning, call on a range of computational tools and programs, and can check steps and solutions.
In mathematics, this makes them qualitatively different from earlier classroom technologies. Students are much less likely to benefit from powerful tools if they do not have the mathematics to judge the outputs, interrogate assumptions, notice errors, compare approaches, and decide whether a result is sensible.
A pragmatic way forward
A pragmatic response is to distinguish between what should remain stable at each stage of schooling and what should be more adaptable.
A stable core would include fundamental concepts, conceptual progressions, reasoning, structure, and the development of mathematical understanding across years of schooling.
A complementary responsive component would include guidance and examples on modelling practices, assessment conditions, and the balance between technology-free, human-led and AI-active mathematical work.
A formal process for a more frequent, progressive review of the mathematics curriculum could be used, with careful attention to continuity and coherence across the stages of schooling. This process would need to be carefully coordinated with active engagement of the teaching profession, and consultation with key stakeholders.
Final thought
How can curriculum remain coherent while becoming more responsive?
This calls for careful thought about what should remain stable; what should be reviewed more frequently; and how assumptions, values and structures should be tested against wider technological and societal change.
Curriculum change needs to account for both future cohorts and current students already moving through an existing sequence of learning.
What mathematical learning do you think should be central for each stage of schooling, and what needs to be reconsidered in the context of AI?
In Part 2, we consider a related question: if AI is now part of the school mathematical landscape, what does it mean to do mathematics effectively?
AI statement:
This blog post was written by a human. ChatGPT was used for final editorial proofreading before being reviewed by a human editor.