Using Large Language Models and AI for SEN

Developments in AI for Education (AIED), particularly around the use of Large Language Models, are providing new vistas and opportunities for the application of educational technology to meet the needs of children with special educational needs. This session will explore, from a critical perspective, cutting edge developments in this area, drawing on a number of case studies of new research and applications in the field.

A key driver in this space is of course the promise of improved personalisation to meet individual needs, that potentially can better support children and young people to achieve more than they could be before. The impetus to develop this technology is clear, and innovation is responding both with specific tailored technologies working alone or very often in tandem with technologies with embedded Large Language Model support. The hope is that the promise of technological advances in AI including LLMs, but also more widely machine learning and natural language processing will augment (although certainly not replace) the role of the teacher. If this technology can encompass and make sense of more data than current technology and the human mind can encompass, it will allow teachers to do more and achieve more in achieving effective inclusive practice for children and young people with SEN.

This session will review the broad evolving technological landscape for the use of AIED for SEN, allowing participants to familiarise themselves with new advances either on the market or emerging to market. It will focus on several example cases of new developments in this area, considering these from a critical perspective. In particular, use will be made of UNESCO’s frameworks for “Human Centered AI” which focuses on how AI can be employed in education in ways which privileges the teacher-student relationship as central to the work of the school. The case study examples will also draw on innovative recent work by researchers in London, Cambridge and Brazil. At the same time, throughout the session, we will draw on the contemporary professional and personal experience of the participants in using AIED and consider applications to their local school contexts.

Text analytics and AI to Support Literacy for SEN

Literacy is of course a key skill for all children as it unlocks the curriculum, and language and literacy development is often a central focus for children with SEN.

AIED and LLM technology is being used in diverse ways to support this – scaffolding, supporting, prompting to provide individualized feedback and support both for reading and writing skills development. Ensuring this can be done whilst avoiding the risk of the “crutch” of dependency is an important element of the implementation of these approaches.

One particular approach is use of text analytics which uses either specific AI technologies or adapted LLMs (or a combination of the two) to analyze students’ writing, identifying particular profiles of strengths and weaknesses for example specific to autistic student writers, and providing feedback possibilities that teachers can then use for feedback and in deciding on next steps in pedagogical approaches for individual writers.

Next Generation Classroom Analysis

Teachers are continually assessing what is going on in their classrooms and making judgements about what works and what does not. Again, new specific AI technologies or adapted LLMs (or again a combination of both) are increasingly being used to give teachers a new radical lens on what is happening in the classroom. Such tools are being used, for example, to analyze transcripts of classroom talk and identify what types of approaches are successful with particular groups of students with particular needs within the classroom, giving the teacher new sources of data on how they could adjust their teaching to be more effective. Similar approaches are also analyzing video to understand students’ emotional reactions during classroom interactions.

The session will involve reflecting on the challenges and opportunities presented by these new technologies for personalisation for SEN and consider practical applications of these technologies emerging on the market. The session will conclude with consideration of how a local needs analysis tool can be used to help participants assess emerging technologies in the SEN space and make informed critical decisions on their application.

Learning Outcomes
Participants will be able to:

• Understand the educational technology landscape for new applications for AIED and LLMs designed for personalisation of learning for children and young people with SEN
• Develop an understanding of the “Human Centred AI” critical model of evaluating and implementing AIED in respect of SEN
• Explore two or more case studies of cutting-edge emerging technologies, including a focus on applications for literacy and language development for SEN, and develop critical skills in assessing such new technologies and making decisions on their use and possible adaptation in their local contexts

Target Audience: Special Educational Needs Officers, Special Educational Needs Teachers, and all educators interested in inclusive practice.