"I'd Never Actually Realized How Big An Impact It Had Until Now": Perspectives of University Students with Disabilities on Generative Artificial Intelligence

Research Background and Issues

• Identified Problems or Challenges:

  • Students in higher education often face barriers related to learning, social interactions, and logistical support, particularly students with disabilities who require additional effort to address these issues.
  • Current university policies and support systems lack adaptability in addressing the impact of generative artificial intelligence (GenAI) on students, especially those with disabilities.
  • The potential impact of GenAI on students' use of accommodations has not been adequately studied.

• Significance of the Issue:

  • According to U.S. data, undergraduate students reporting disabilities account for approximately one-fifth of the total student population, and this group is steadily growing.
  • Generative AI has the potential to reduce educational inequities by improving accessibility (e.g., personalized learning, communication support, self-advocacy).
  • Without proper policy guidance, AI technologies may exacerbate biases, such as perpetuating stereotypes about individuals with disabilities.

• Research Motivation and Related Work:

  • Previous research has explored some barriers faced by students with disabilities in higher education, but there is limited study on the practical application of GenAI in supporting these students.
  • The authors propose that generative AI could serve as a potential tool for improving resource allocation through personalization and accessibility, but its risks (e.g., inaccurate outputs and implicit biases) also warrant attention.
  • This study addresses the research gap between AI and accessibility in higher education by collecting data from students with disabilities.

Solutions

• Methods or Solutions:

  • This study employs a mixed-methods approach, distributing an online survey to 62 students at two U.S. universities and conducting semi-structured interviews with 21 interested participants.
  • The research team utilized surveys, interviews, and qualitative analysis to explore how GenAI tools influence university students' learning experiences, accessibility, and use of accommodations in both academic and non-academic contexts.

• Innovative Aspects:

  • A deep investigation into the unique impact of generative AI on the academic and social activities of students with disabilities, revealing rich details from students' self-expression and perceived impacts.
  • Exploration of the potential for GenAI applications in university teaching and policy, with a particular emphasis on supportiveness and inclusivity.

• Implementation Steps:

  1. Data Collection Phase: Gather students' experiences with GenAI tools through questionnaires and interviews.
  2. Data Analysis Phase: Decode students' usage of GenAI tools and their perspectives on support systems through statistical and qualitative coding.
  3. Recommendations Output: Based on findings, propose recommendations for GenAI educational resources, university policies, and the development of accessible technologies.

Research Outcomes

• Specific Findings:

  • GenAI tools are primarily used for course support (e.g., writing assistance, information retrieval), academic research (e.g., code debugging, literature reviews), and daily life tools (e.g., time management, self-care).
  • Historically under-supported groups (e.g., individuals with psychological disabilities and ADHD) reported effective assistance from these tools, such as personalized learning content and life routine planning.

• Advantages Compared to Existing Solutions:

  • Due to its accessibility (e.g., 24/7 online support) and low entry barriers, GenAI is viewed by many students with disabilities as a complementary alternative to formal support systems (e.g., Disability Services Offices, office hours).
  • GenAI tools significantly reduce the administrative workload required for students to access support, helping to generate tailored communication plans.

• Experimental or Evaluation Results:

  • Most participants believed they had the knowledge and resources to use GenAI tools and planned to continue exploring them.
  • Nevertheless, concerns about AI inaccuracies and potential biases were widely mentioned, indicating the need to enhance students' critical thinking regarding AI outputs.

• Limitations and Future Directions:

  • Limitations:
    1. The sample is limited to two specific public universities, which may not be generalizable to other educational settings.
    2. Most participants were familiar with only a few tools (e.g., ChatGPT) and lacked in-depth experience with other widely used GenAI tools.
  • Future Directions:
    1. Investigate how individuals with undisclosed disabilities utilize GenAI tools to design more compatible solutions for these students.
    2. Explore the feasibility of formally integrating AI educational resources and training into university policies.
    3. Conduct longitudinal studies to analyze changes in the overall impact of GenAI on students' lives.