A Study of the Use of Artificial Intelligence in Differentiated Instruction

The study builds an AI support framework based on Tomlinson's theory, and achieves accurate learning diagnosis through five-dimensional data (cognitive, behavioural, emotional, environmental and social). Practice has proven that AI can dynamically optimise teaching. However, there are challenges in AI application: over-reliance on technology undermines teacher-student interaction, algorithmic bias exacerbates educational injustice, and data privacy risks need to be regulated. The study emphasises that AI in education should be based on the essence of “educating people”, and promote the role of teachers to shift from “knowledge transfer” to “growth guidance” under the ethical framework, and synergise technological empowerment and humanistic care. The future needs to explore localised adaptation paths. In the future, it is necessary to explore the local adaptation path, not only to make use of AI to break through the bottleneck of large-scale education but also to prevent the alienation of the technology, and build an intelligent education ecology of “teaching in accordance with one's aptitude”.