Â鶹ÒùÔº

This report examines how generative artificial intelligence (AI) is used as a possible form of assistive technology by students with special educational needs and disabilities (SEND) in secondary mathematics. Drawing on a small-scale practitioner inquiry involving two students with SEND, the study analysed when AI use supported understanding and when AI-supported interaction broke down.

Key findings include:

• AI supported learning when it was used to clarify specific points of uncertainty, but hindered learning when explanations were too dense, too fast paced or misaligned with prior knowledge.

• Breakdowns in understanding were not usually resolved by AI alone; they required teacher mediation, particularly to reconnect explanations to familiar methods and carry important ideas across a sequence of work.

Implications for policy and practice include:

• AI should not be treated as a substitute tutor, but as part of an interaction between student, teacher and technology.

• Teachers should explicitly teach students how to use AI, and schools should prioritise supported access and teacher development.

• Schools should ensure that students with SEND have supported access to AI tools, rather than being excluded from them because of concerns about misuse or risk. 

• Future AI design should pay closer attention to enabling step-by-step interaction, addressing specific points of difficulty and supporting active engagement over time.