How To Use AI For Maths: A Practical Guide For Teachers And Heads of Department

AI for maths is knocking on the classroom door. But for most of us, especially those in a leadership role in schools, figuring out how to use these maths AI tools effectively and safely feels like navigating without a map.

In this guide, I’ll share practical strategies and key dos and don’ts for using AI in maths, based on what I’ve found works — and what doesn’t.

This blog is part of Third Space Learning’s AI in education series and uses some examples drawn from Skye, a voice-based AI maths tutor. 

Why AI matters for today’s maths classroom

When we talk about “AI in maths,” what do we actually mean?

It’s a broad church. Tools range from advanced AI models such as general-purpose chatbots like ChatGPT that solve equations and generate practice problems, to dedicated platforms like TeachMateAI, which create maths-specific resources and answer complex equations.

Beyond that, we are starting to see a future of text and voice-based AI math tutors promising engaging and tailored content or support for individual student needs.

What’s clear is that just as the tools vary so do the outputs. So if you’re a member of the SLT or a curriculum or technology director looking to drive the uptake of AI maths tools in your school or trust, you’ll need a good understanding of the options available before in order to assess if the maths AI actually address the problem you’re trying to solve.

​​Potential benefits and new opportunities from AI in maths:

• Instant solutions or explanations of tricky topics: Learners stuck at home can get immediate help from an online maths solver, reducing frustration and making study sessions more productive.
• Time-saving for educators: AI helps to speed up drafting differentiated maths problems, worked examples, or quizzes, ensuring reliability in the accuracy and quality of the materials provided.
• Potential for personalised practice: Maths tools might generate tailored problems and calculations, or provide on-demand step-by-step explanations for complex math problems.
• Exam preparation: Maths AI solvers can assist in creating practice papers and spaced reminders for targeted exam preparation.

Significant cautions of AI in maths:

• Inaccuracies: General text or voice-based chatbots can confidently get maths wrong (reasoning errors, simple calculations, particularly times tables above 12×12 and misinterpretations over context). Relying on unchecked output is risky when using it to solve problems.
• Limited functionality: AI tools have limited capabilities in accurately solving complex maths problems. This can lead to errors in higher-level calculations and nuanced problem-solving.
• Over-reliance: Students might lean too heavily on AI maths, bypassing thinking and problem-solving. They may struggle with resilience and not independently solve tricky algebra or geometry math problems.
• Data privacy concerns: Many tools collect data. Understanding what is collected, how it’s used, and where it’s stored is paramount under data protection regulations like GDPR.

How AI maths supports learning and personalised instruction

So what is AI in maths useful for, and how can you or your teachers best integrate it into your work? Here are a few examples.

Quick explanations for homework help

A student is stuck on a specific step in a prime factorisation problem at home. Instead of asking a math solver for the answer, they can upload a clear photo of the problem to a vetted AI tool and ask for an explanation of the method or explain why a particular rule applies.

You may need to model or provide a handout to show how to ask the AI for a method explanation or a concept clarification, rather than just give the final answer.

Generating differentiated examples with caution 

You need a few extra examples of finding the area of compound shapes for an intervention group struggling with a specific type of shape combination.

They could use AI to generate a starting list of shape descriptions or diagrams. Then, they carefully edit and select problems that ensure appropriate challenge and variation and fit the group’s:

• Specific needs;
• Common misconceptions;
• learning objective.

Checking understanding and identifying misconceptions

After a lesson on solving linear equations, students use a supervised AI maths tool to solve one or two practice problems. The task isn’t just to get the answer, but to ask the AI to explain each step in its solution.

To help surface any misconceptions and address input problems, students compare the AI’s explanation to their work and discuss any discrepancies or unclear steps with a partner or the teacher.

Exploring alternative methods

Your teachers use a maths solver on the interactive whiteboard to quickly show students multiple ways to solve a problem. For example, students can drag, paste, or upload images of their math problems into the solver area, making it easy to solve a quadratic equation by factorising, using the quadratic formula, or graphically.

The focus is on comparing the efficiency and logic of the different methods, not just the final answer.

Generating real-world contexts

You need a quick, real-world problem involving percentages for a lesson starter. They could ask an AI to generate a scenario from photos of worksheets or images of math equations, which they then adapt and refine to make it relevant and mathematically sound. These scenarios often come as fully worked examples that explicitly ask learners to solve word problems tied to everyday life.

AI maths tutoring

A group of students need extra maths support in addition to whole-class quality-first teaching. Your school doesn’t have the budget or staff resources to implement one-to-one or small group tutoring sessions, or online tutoring. Instead, many schools like yours are increasingly turning to personalised, conversational AI maths tutoring over traditional tutoring to support the learners who need it.

Key maths AI considerations for SLT and Heads of Maths

Before implementing AI for maths, or any other subject, school leaders need a coherent strategy for AI use and clear guidance for staff and students to follow. Without this, it’s hard to ensure AI is used carefully, consistently and safely. 

Safeguarding and student wellbeing

Schools must ensure a safe environment and mitigate potential risks when interacting with AI tools with chat functionalities, either text or voice-based. 

Any school-based AI decision should be based on the UK Government’s Generative AI: Product Safety Expectations.

Here are some further ideas to add to the discussion: 

• Supervised use: Monitor usage and interactions closely in school, especially for younger students, and intervene as needed.
• Avoid unsupervised chat: Often, parental consent is required for students aged 13-18. Unsupervised interactions can present unpredictable safeguarding risks.
• Humans-in-the-loop: Consider educational tools where educators can oversee and review student progress and interactions. Additionally, consider AI tools that use human expertise to provide maths content for students.  

Academic integrity and cheating

Perhaps the most pressing fear for many users is the potential misuse of maths solvers. AI maths solvers can provide instant solutions by solving equations and showing step-by-step solutions. Students can copy solutions without doing any heavy cognitive lifting to understand the math concept.

Students can get 100% on maths homework tasks but learn nothing. There are no easy ways around this (except for the total ban of AI and a retreat to handwritten homework).

Acknowledge the risk

Uploading homework is a common practice among students, and simply banning solvers is unlikely to work; students will find and use them outside of school. Instead, acknowledge the risk openly and build strategies to mitigate it.

Require working and reflection

Insist that students show their working out, even when using an AI maths solver. Furthermore, encourage them to explain the steps or the logic behind the solution provided by the math solver in their own words.

The “flipped perspective”

Instead of using maths AI to get the answer to that algebra question, train students to check their answer after they’ve attempted it, or to understand a step they got wrong when solving math problems. 

Supervised environments

For high-stakes tasks or when introducing AI for maths problem-solving, use it in a supervised classroom setting to ensure students use it appropriately.

Always ask why

While we cannot stop students from using AI at home, we are in control at school. Before allowing them to use age-appropriate AI tools consider ‘why’ you want them to use it and how it will help them.

Curriculum alignment 

Different AI tools have different understandings of the national curriculum, exam board specifications, teaching strategies, sequences of math concepts and pedagogies.

An AI maths tool trained on a vast dataset might have a general understanding of maths, but won’t automatically know the:

• Specific order you teach topics;
• Methods and strategies you emphasise;
• Common misconceptions specific to your students.

So, how can you ensure your prompts provide a useful output?

Verify content

If using AI to generate maths practice problems, explanations, or lesson materials, you must verify that the content is accurate and aligns with your school’s curriculum or syllabus such as NCETM or White Rose, teaching sequence, exam board specifications (e.g, Edexcel or AQA) and the national frameworks.

Age-appropriateness

AI may give a mathematically correct explanation, but it might be delivered in an overly complex way. Ensure that the AI provides a detailed explanation that is age and ability appropriate. Adapt or reject AI outputs to suit your learners.

Set boundaries

Decide what maths AI can be used for. A basic arithmetic practice generator for creating similar problems might be low risk, while an AI explaining complex calculus concepts might be higher risk for inaccuracies or complexity.

Teacher Knowledge and CPD

Expecting teachers to navigate the complexities of AI tools, master their potential, identify their pitfalls, and guide students effectively, all without support, is unreasonable. It will lead to frustration and inconsistent application.

• Basic training: Staff need basic training on how common AI math solvers and other tools work, and where they can go wrong. 
• Integrate CPD: Implement short, practical CPD sessions into existing structures. This could be a quick demonstration and discussion in a staff meeting, or a shared resource bank of vetted tools and prompt examples.
• Build confidence: Equip staff to confidently explore and use AI as a support for maths. Start with a departmental or phase meeting to give colleagues a chance to experiment with AI in a safe environment and provide an opportunity to share resources. 

The Teacher Prompts newsletter is a free weekly bite-sized newsletter that keeps teachers up-to-date on AI and provides ideas and prompt suggestions across all subjects and phases. 

Data protection and privacy

This is non-negotiable. Any online tool students use, particularly within a school context, must comply with data protection regulations and GDPR in the UK/EU and the UK Government guidelines.  

Schools have a legal and ethical responsibility to protect student information.

Read privacy policies

School leaders must read and understand the privacy policy of any AI tool being considered for use with students. 

Avoid sharing sensitive data

Ensure that teachers and students are not inadvertently sharing sensitive personal information when using AI tools, especially general chatbots not designed for education. 

Vet tools

Prioritise educational tools built with privacy and safeguarding in mind, and that are transparent about their data practices.

Any school using AI should have an AI policy. Third Space Learning’s safeguarding leads have created an AI policy template for the use of Skye, the conversational AI maths tutor, and other AI tools. You can download the editable AI policy template for free and adapt to your school.

Equity and access

Many bullish AI commentators heavily promote AI as the great leveller of educational outcomes. Not many are concerned that the rise of AI may actually increase the attainment gap. 

Reliance on AI math solvers, especially for homework or independent study, may exacerbate existing digital divide inequalities. Not all students have reliable home internet, so setting homework that requires AI maths tool use could disadvantage some learners. 

If integrating AI into your teaching strategies, you must plan for equitable access. This might include structured computer sessions for AI-assisted homework or implementing AI tutoring sessions during the school day.

Practical dos and don’ts

Let’s translate these considerations into actionable steps for the classroom and department. Use these as quick checks before introducing or recommending a tool, ensuring the interface is user-friendly and accessible.

Do: Prioritise deep thinking over just getting answers.

Don’t: Allow reliance on final answers; require working/explanation for academic integrity.

Insight: Under 13s shouldn’t use ‘open’ chatbots in school; secondary students need consent.

---

Do: Implement clear safeguarding and check data policies thoroughly.

Don’t: Allow unsupervised, unfiltered chatbot use for young/vulnerable students.

Insight: Make sure any AI tool you use has a vetting process. Many don’t, watch out for this.

---

Do: Ensure AI outputs align with your curriculum and pedagogy. Specify your schools frameworks.

Don’t: Assume AI suggestions are correct or appropriate; use professional judgment.

Insight: AI lacks pedagogical context (e.g., Year 3 interest problems); adapt any AI outputs heavily to match your school context.

---

Do: Provide staff training and CPD on AI maths tools.

Don’t: Expect teachers to figure AI out alone; encourage knowledge sharing and support.

Insight: Group experimentation builds comfort and leads to useful discussions.

---

Do: Plan for equity and access; provide alternatives for students lacking home resources.

Don’t: Disadvantage pupils without connectivity; bridge digital divides.

Insight: AI tutoring tools provide personalised maths support at a fraction of the cost of traditional tutoring.

---

Do: Guide pupils on verifying AI accuracy; teach critical thinking.

Don’t: Assume AI solutions are always correct or use the only method; develop critical evaluation.

Insight: Demonstrate AI errors; it’s a tool requiring checks.

---

Do: Integrate AI strategically for specific, high-value tasks such as checks, examples, and exploring methods.

Don’t: Let AI replace teacher-led instruction or essential practice.

Insight: Chunk tasks for AI; don’t ask for whole lesson plans.

---

Common pitfalls and how to address them

Beyond the overarching dos and don’ts, let’s touch on a few specific pitfalls that can hinder productive use of AI tools for teachers and students in maths:

Instant worksheet trap

AI quickly generates questions but often lacks pedagogical structure, including scaffolding, variation and addressing misconceptions.

For example, when creating a year 3 fractions worksheet, AI might include improper fractions and lack of concrete-pictoral-abstract support.

Solution: Use AI for inspiration or ideas, but build core resources based on sound pedagogy, requiring significant teacher editing and structuring.

Shallow understanding

Students get answers without grasping conceptual understanding, and reproduce steps without internalising reasoning.

For example, if asking an AI maths solver to complete 3(2x − 5) = 4x + 7 the learner will not understand the method.

Solution: Require explanation, reflection, and discussion. Frame AI as a tool for exploring methods and processes, not just getting the answer.

Teacher workload 

While helping to reduce teacher workload in the long term, initial AI integration can increase workload, particularly for those involved in vetting, verifying and strategising.

Solution: Start small, pilot one tool for a specific purpose, share the load, and focus on areas where AI genuinely saves time without compromising quality.

Human teachers are irreplaceable

Let’s be absolutely clear: AI is a tool, but it cannot replicate the truly essential, human aspects of being a maths teacher. Used wisely, maths AI becomes an assistant, not a substitute. AI alone cannot:

• Spot subtle misconceptions through observation and interaction.
• Provide emotional support, encouragement, and build resilience.
• Adapt feedback in the moment based on student context and state.
• Build positive, trusting relationships for a safe learning environment.
• Curate the learning journey with deep pedagogical insight, linking ideas and igniting curiosity.
• Inspire a love for mathematics through personal passion.

Your expertise, intuition, ability to connect with students, understand their individual needs beyond just their mathematical output, and orchestrate their learning experience are irreplaceable. 

AI should be viewed as a sophisticated assistant, a digital calculator on steroids, a resource generator, or a different perspective on a problem, but never a substitute for your vital human role in the classroom.

 

Talking to students about AI

The final, but very important element for using AI in maths or any other subject is the students. Find out what they already know or think about AI. How are they using AI in maths already, and how can you ensure that it is a responsible and effective use as part of their learning? It’s crucial to have honest communication with students about AI is crucial so you can help them understand its potential as a tool and its limitations. Here are some specific guidelines:

• Verify AI: Teach students AI can be wrong; verify outputs with reliable sources, or develop critical thinking.
• Process over answer: Explain copying stunts learning; value is in the process. Use AI to understand and check, not bypass thinking.
• AI as a partner: Encourage using AI proactively, explain the steps, generate practice and explore methods.
• Basic e-safety: Remind students of online safety, no personal details, supervised use, and appropriate interactions.

For example before any student starts a lesson with Skye Third Space Learning’s voice-based AI maths tutor, they are taught how to communicate with her safely and effectively to get the most out of every session.

Next steps for AI maths tool use in schools 

The age of AI in education is here. While it brings exciting possibilities for supporting teaching and learning and solving math problems, it also demands careful consideration, open discussion, and a pragmatic approach, especially in a foundational subject like maths.

Ignoring it isn’t an option, but neither is blindly adopting every new tool that appears.

For heads of maths and school leaders, the key takeaways are clear:

• Prioritise safeguarding
• Protect academic integrity
• Ensure curriculum alignment
• Invest in teacher CPD
• Address data and equity

Start small, perhaps piloting one tool for a specific purpose with a supportive group of teachers. Gather feedback from both staff and students, discuss what works and what doesn’t, and refine your approach and school policy as you learn.

AI is making huge advancements every day. We must continuously adapt our strategies and approaches to AI use in maths and wider education to support the vital work of teachers, not to replace it.

FAQs