The rapid rise of generative AI (GenAI) tools like ChatGPT, Microsoft Copilot, and Google Gemini has sparked heated debates in educational circles. From concerns about academic integrity to excitement over their potential to revolutionise learning, these tools are reshaping both the academic and business landscapes. A recent survey conducted at the University of Edinburgh’s School of Mathematics sheds light on how mathematics students are navigating this AI-infused terrain and what their perceptions reveal about the future of education. The results were published as a Feature in the London Mathematical Society Newsletter, Issue 513, Dec. 2024.
The State of GenAI Adoption Among Mathematics Students
The findings reveal that an overwhelming majority of students—93%—have experimented with GenAI tools. These tools are being used for tasks ranging from solving technical problems to explaining complex concepts and even for casual exploration. Interestingly, ChatGPT emerged as the most widely used platform, with others like Claude, Gemini, and Perplexity trailing behind.
This widespread adoption is no surprise. Mathematics students–like those of other STEM subjects–are often at the forefront of technological adaptation and adoption, find these tools invaluable for quick problem-solving and conceptual understanding. However, the survey also unveiled a dichotomy in how students perceive the role of AI in education.
The Benefits: Efficiency, Accessibility, and Future Readiness
Students appreciated the immediacy and accessibility of GenAI. Over half of the respondents felt these tools saved time, provided personalised feedback, and enhanced writing efficiency. The survey underscores a growing belief that proficiency in GenAI will be crucial for future employability, a sentiment echoed by employers who increasingly prioritize AI literacy.
Moreover, the potential for AI to serve as a personalized tutor—offering guidance tailored to individual learning styles—is an exciting frontier. This capability could democratise education, making high-quality resources available to a broader audience, especially in resource-constrained settings.
The Drawbacks: Trust, Accuracy, and Educational Value
Despite these advantages, skepticism persists. A majority of students (85%) acknowledged that GenAI outputs often lack reliability and contextual appropriateness. This lack of trust is a critical barrier to wider adoption, as students noted the frequent inaccuracies in AI-generated answers. Interestingly, those less knowledgeable about a concept were more likely to overestimate the accuracy of GenAI, a pattern also observed in studies of other STEM fields.
Students expressed a clear preference for human instructors, with 75% agreeing that humans are more effective than AI in guiding coursework. This preference aligns with findings from a study (cited by the authors of he study) comparing Japanese students learning English with either a human tutor or a chatbot. Over 15 weeks, students with human tutors reported an increased interest in learning, while those paired with chatbots showed a decline. Similarly, the Edinburgh survey highlighted that GenAI tools might enhance efficiency but fall short of fostering deeper engagement or sustained learning.
Confidence and Essentiality of Tools
When it came to confidence in using GenAI tools, 57% of students felt equipped to use them effectively. However, a notable 57% also indicated that these tools were not essential to their learning journey, suggesting that while GenAI is helpful, it is not indispensable. Additionally, 63% of respondents preferred traditional search engines for information retrieval, further emphasizing a cautious reliance on AI.
Interestingly, students’ ambivalence about the tools’ essentiality reflects a broader uncertainty about their long-term educational value. While many saw the potential of GenAI for saving time or simplifying tasks, there was a shared concern that overreliance on these tools might undermine critical thinking and foundational skill development.
Implications for Teaching and Assessment
For educators, these findings pose a significant challenge: how to integrate GenAI effectively without compromising learning outcomes. The answer might lie in rethinking assessment methods. If traditional homework and exams become less indicative of student understanding due to AI use, alternative methods like project-based assessments or oral exams could fill the gap.
Moreover, fostering a critical approach to AI tools is essential. Teaching students how to evaluate the reliability of AI-generated outputs and encouraging them to view AI as a complement rather than a substitute for traditional learning could mitigate some of the risks.
The Bigger Picture: Preparing for an AI-Driven World
The survey highlights a critical need for higher education to evolve in tandem with technological advancements. While GenAI tools are not a panacea, their potential to augment learning is undeniable. Universities must strike a balance between leveraging these tools and preserving the foundational skills that underpin disciplines like mathematics.
As GenAI continues to permeate academia and the workplace, educators, policymakers, and students must engage in an ongoing dialogue about its role. The University of Edinburgh’s proactive approach—researching and addressing these challenges head-on—is a model worth emulating.
For students, educators, and stakeholders in education, the message is clear: GenAI is here to stay. Embracing its benefits while remaining vigilant about its limitations will be the key to unlocking its full potential. Let us view this technological wave not as a threat but as an opportunity to redefine and enrich the learning experience.
As we move forward, let this serve as a reminder: the essence of education lies not just in mastering tools but in cultivating critical thinking, creativity, and the ability to adapt—qualities that no AI can replicate. Here are some thoughts on the benefits and pitfalls of its use in mathematics eduacation
The Bright Side
- Personalized Learning Paths: Generative AI can analyse a student’s strengths and weaknesses, adapting instruction to individual needs. For example, tools like AI-driven tutoring platforms can create tailored problem sets, ensuring students receive practice where they need it most.
- Immediate Feedback: AI systems provide instant feedback on problem-solving attempts, enabling students to learn from mistakes in real time. This can reduce frustration and accelerate the learning curve.
- Enhanced Problem Solving and Visualisation: Generative AI can illustrate complex mathematical concepts using interactive graphs, 3D models, or stepwise solutions. This helps students grasp abstract ideas more effectively.
- Efficiency for Educators: Teachers can leverage AI to automate grading, create assignments, and even develop lesson plans. This reduces administrative burdens, allowing more time for one-on-one student interaction.
- Bridging Knowledge Gaps: AI tools can be a valuable resource for students who lack access to quality education, offering a consistent learning companion that doesn’t require constant human oversight.
The Dark Side
- Overreliance and Reduced Critical Thinking: Students may become dependent on AI for answers, neglecting the development of problem-solving and reasoning skills. Blindly following AI solutions can lead to a superficial understanding of mathematics.
- Equity and Accessibility Issues: Advanced AI tools require robust internet access and expensive devices, potentially widening the digital divide. Students in underprivileged areas may find themselves at a disadvantage.
- Ethical Concerns: AI systems can inadvertently promote academic dishonesty by enabling effortless cheating. This undermines the integrity of education and assessment processes.
- Inaccuracies and Misinterpretation: While AI is powerful, it isn’t infallible. Generative models may produce incorrect solutions or explanations, confusing learners and leading to misinformation.
- Reduced Teacher Involvement: Over-dependence on AI tools might lead to a diminished role for educators, weakening the mentor-mentee relationship that is crucial for holistic learning.
Navigating the Balance: Best Practices for Using AI in Math Education
To harness the benefits of generative AI while mitigating its drawbacks, a balanced approach is essential:
- Complement, Don’t Replace: Use AI as a supplement to traditional teaching methods. Encourage students to engage with the technology while maintaining critical thinking and independent problem-solving skills.
- Teacher Training: Equip educators with the knowledge to effectively integrate AI tools into their teaching. Teachers should guide students in responsibly using AI for learning, not shortcuts.
- Develop Ethical Guidelines: Institutions should establish clear policies on AI usage to prevent misuse and ensure academic integrity.
- Focus on Equity: Ensure AI resources are accessible to all students, regardless of socioeconomic background, by providing affordable or free tools.
- Critical Evaluation of AI Outputs: Teach students to verify AI-generated solutions and understand underlying principles, fostering a deeper grasp of mathematics.
Generative AI has the potential to revolutionise mathematics education, making it more engaging, personalised, and efficient. However, its unmoderated use poses significant risks to critical thinking, equity, and ethical practices. Educators, policymakers, and technologists must work collaboratively to maximise AI’s benefits while addressing its challenges. When wielded wisely, this double-edged sword can be a transformative force in shaping the mathematicians of tomorrow.