Table 1 Mapping the impact of AI implementation in the learning design

Individualised learning, achievement goals and expectations

①

Task demand and cognitive quality matching

Implement intelligent algorithms to analyse individual student learning patterns and dynamically adjust the task demands and cognitive quality of instruction in real-time.

Tailored learning experiences that align with each student’s proficiency level, minimising frustration and enhancing engagement.

+ A student excelling in calculus faces more advanced problems, while another struggling with basic algebra receives simplified tasks. This tailored approach aligns with each student’s proficiency level, minimising frustration and enhancing engagement by providing challenges that are appropriately matched to their abilities.

− The AI platform struggles to accurately assess individual learning patterns. As a result, a student ready for multiplication is confronted with overly challenging division problems, causing confusion and frustration. Conversely, a student comfortable with multiplication tasks is presented with basic addition, leading to boredom.

②

Agency and cooperation in learning environments

Incorporate AI-supported adaptive learning platforms that allow students to progress at their own pace, fostering autonomy. Integrate collaborative AI tools to encourage cooperative problem-solving.

Empowered students with a sense of control over their learning journey and enhanced social interactions that positively influence their emotional experiences.

+ In an AI-supported maths platform, a high-performing student working on algebra receives dynamically adjusted tasks. If they excel in solving linear equations, the system challenges them with quadratic equations or introduces more complex variables.

− The AI tools create confusion in coordinating efforts among students working on a trigonometry project, leading to unproductive teamwork and hindering a meaningful understanding of the topic. This diminishes the sense of autonomy and positive social interactions among students.

Individualised achievement goal structures and expectations

Leverage predictive analytics to understand individual student capabilities and set personalised achievement goals. AI-driven recommendation systems can guide students in setting realistic expectations.

Students work towards achievable goals tailored to their abilities, reducing anxiety and fostering a sense of accomplishment.

+ An AI-driven recommendation system suggests realistic targets for a student studying calculus. The student, with guidance, breaks down complex integration problems into manageable steps, achieving goals aligned with their abilities. This approach reduces anxiety, enhances focus, and fosters a sense of accomplishment in mastering challenging mathematical concepts.

− A struggling student in algebra receives recommendations for advanced calculus objectives. This mismatch creates undue stress, hindering the learning process and leading to feelings of inadequacy. The lack of accurate goal-setting guidance diminishes the positive impact on student motivation and accomplishment in mathematics.

③

Improved value induction

Use machine learning algorithms to assess students’ preferences and interests, delivering content that resonates with their individual values.

Increased intrinsic motivation as students find personal relevance in mathematical concepts, leading to a positive emotional response.

+ A student interested in architecture explores geometry through designing structures. This personalised approach increases intrinsic motivation, as the student finds direct relevance and personal connection to mathematical concepts, resulting in a positive emotional response and heightened engagement with the subject.

− A student passionate about statistics receives content focused on geometry puzzles, creating a disconnect. This mismatch diminishes intrinsic motivation, as students fail to find personal relevance in the assigned tasks, leading to a negative emotional response and reduced engagement with mathematical concepts.

Moderating feedback mechanisms

④

Feedback strategies to preserve sense of control

Implement AI-powered feedback mechanisms that provide constructive guidance (adaptive feedback) without undermining students’ confidence.

Constructive feedback that encourages improvement without eroding students’ or teachers’ sense of control, promoting a positive emotional response to challenges.

+ A student working on a geometry-based task receives feedback from AI that highlights their areas of growth since they last worked on a similar activity 6 months previously. The ability to frame the feedback within the larger picture of overall growth enhances their sense of achievement and capability and gives them a perception of greater control over future tasks.

− A student working on algebraic expressions receives generic feedback that doesn’t address specific errors or strengths. The lack of precision in feedback diminishes the sense of control, as students struggle to understand and improve.

Emotion regulation

⑤

Retraining attributions for failure and success

Develop AI-driven interventions that help students reframe perceptions of failure and success through personalised feedback and motivational prompts.

Shift in mindset as students learn to view setbacks as opportunities for learning, reducing anxiety associated with performance.

+ If a student struggles with a challenging statistics problem, the system offers specific guidance on improvement and highlights the effort invested. Motivational prompts encourage viewing setbacks as opportunities for learning.

− A student facing challenges in geometry receives generic messages that overlook the deliberate decision by the teacher to allow the student to persevere through difficulties. The lack of understanding and acknowledgement of this intentional teaching strategy hinders the development of skills related to perseverance and independent learning. There is a concern that the reliance on generic feedback may lead students to seek quick fixes and undermine their ability to navigate challenges independently, raising questions about the sophistication of the learning analytics employed.

⑥

Fostering self-regulation of emotions

Implement emotion-aware AI systems that recognise and respond to students’ emotional states. Integrate virtual mentors or chatbots to offer support and strategies for emotional self-regulation.

Enhanced emotional regulation, as students learn to identify and manage their emotions, contributing to a more positive learning experience in mathematics.

+ An emotion-aware AI system in maths recognise when a student feels frustrated while working on a challenging algebra problem. A virtual mentor engages the student, providing supportive feedback and suggesting effective strategies for emotional self-regulation, such as taking short breaks or breaking down the problem into smaller parts.

− Despite the implementation of emotion-aware AI systems, the recognition of students’ emotional states in maths proves inconsistent. For example, a student struggling with fractions receives no targeted emotional support, as the system fails to detect signs of frustration. The absence of timely interventions hampers emotional self-regulation, leaving the student feeling overwhelmed and negatively impacting their learning experience in mathematics. The limited effectiveness raises concerns about the overall support provided by virtual mentors or chatbots.