Umeå universitets logga

The rapid integration of artificial intelligence (AI) into society is widely regarded as offering both unprecedented opportunities and significant challenges for education systems worldwide. Schools, entrusted with a vital compensatory mission, are now responsible for addressing educational inequities while equipping students with essential skills for an AI-driven world. This mission ensures that all students, regardless of socio-economic background, functional abilities, or ethnic origin, have equal opportunities to succeed. However, educational inequalities are widening due to factors such as residential segregation and school-choice reforms, which have led to increasing socio-economic disparities between schools. In Sweden, these challenges are evident in national statistics, with 16% of students failing to qualify for upper secondary education in 2023 (SNAE, 2024). As AI becomes more embedded in education, its potential role in promoting equity and inclusion must be critically examined.

This study explores the perceptions, experiences, and future visions of first-year special education teacher students regarding the integration of large language models (LLMs) into their professional practice. Using a mixed-methods approach, data were collected through a survey featuring 30 Likert-scale statements and seven open-ended questions. The quantitative data captured participants' familiarity with AI tools, their perceived opportunities and challenges, and their aspirations for AI integration in education. Qualitative responses provided deeper insights into specific applications, ethical considerations, and strategies for leveraging LLMs in special education. Hence, AI and digital tools are frequently presented as potential solutions for bridging educational divides, offering new ways to support communication, literacy, and self-regulation. However, findings from this study reveal that while participants recognize the potential of LLMs for supporting personalized learning and inclusive practices, they also express concerns regarding ethical dilemmas, data privacy, and the lack of professional training required for effective AI implementation.

In the Nordic educational context, where equity and inclusion are fundamental principles, AI presents both opportunities and risks. When applied thoughtfully, AI has the potential to reduce educational disparities, enhance literacy, communication, and self-regulation skills, and mitigate socio-economic and functional inequalities. However, without adequate support for students with limited technology skills, reading and writing difficulties, or intellectual challenges, AI may instead exacerbate educational inequality, disproportionately benefiting students who already possess strong digital competencies while leaving others further behind.