Neurodivergent students are enrolling in higher education at increasing rates internationally. Therefore, it is important to understand how new educational innovations can be used to support neurodivergent learners. Once such innovation is artificial intelligence (AI). AI programs have proliferated across higher education in recent years, especially with the advent of ChatGPT. AI tools can be beneficial for both students and instructors. On the student end, AI tools can personalize the learning experience and provide quick, individualized feedback, yielding positive effects on engagement and academic achievement. An individualized, responsive learning environment represents an ideal situation in which neurodivergent learners can thrive. As a result, the popular press has emphasized the potential for AI tools to revolutionize education for neurodivergent students. This, in turn, has led to a notable discrepancy between how the public views the impact of AI tools for neurodivergent students, and the actual state of the research.
The Reality: A Key Research Gap
The reality is that very few studies of AI in higher education include neurodivergent students and, of those that do, the majority are focused on tool development rather than evaluating the effectiveness of existing tools for neurodivergent students. Furthermore, even when considering the few studies that have examined effectiveness, methodology is usually not rigorous and accessibility is not usually considered. This research gap has created an environment in which the some of the most innovative educational technology may only be effective for neurotypical students.
Rethinking AI: From Savior to Advocate
While this identified research gap is concerning enough on its own, it is also worth discussing how AI studies including neurodivergent students are framed. Neurodivergent students are oftentimes framed negatively – as a societal problem to be solved, or a burden that can be lifted by AI. This, in turn, creates the impression that AI is a powerful savior, relieving the burden placed on parents and teachers. Instead of being awed by AI – the powerful, infallible solution, we should give our respect to the students themselves. By partnering with neurodivergent students and their families, we can conduct research that is focused on their experiences and how AI can help them thrive. This also means that, rather than viewing AI as the infallible hero, we must accept its shortcomings and the fact that it may not being living up to the hype that is being fed to the public.
Building an Inclusive Future for AI
To address these research gaps, it is first important to determine which AI tools that neurodivergent students are currently using, the purposes for which they are used, and the frequency with which they are being used. Once these foundational questions are answered, the focus can shift to assessing the effectiveness of these tools for neurodivergent students while also addressing potential accessibility challenges. By partnering with neurodivergent students and learning about their first-hand experiences with AI, they become co-designers and self-advocates in the process of developing AI tools that are effective for neurodiverse users. Additionally, the goals of utilizing AI in higher education should be reframed from outperforming human capabilities to the potential for AI to empower all students.
In addition to exploring how neurodiverse students in higher education interact with AI tools, it would be valuable to explore whether AI tools can assist instructors to ensure that their curricula are as accessible and effective as possible by integrating Universal Design for Learning (UDL) principles that allow all students, regardless of neurodivergent status, to engage with and respond to course materials. Important components of UDL include designing multiple means of engagement, representation, and expression to optimize the learning experience for all students.
ChatGPT is one popular genAI program that has proliferated across higher education institutions in recent years. One feature of ChatGPT is the ability to create custom GPTs that are tailored to specific tasks. The custom GPT is configured by providing it with novel instructions and specialized knowledge, thereby increasing its effectiveness at a given task. By configuring a custom GPT with information and instructions about the implementation of UDL principles, it will be able to assist instructors with modifying course materials and/or curricula to ensure that they are applying the UDL framework to their courses. For example, an instructor could upload lecture materials to the custom GPT and receive feedback on its adherence to UDL principles, as well as how the lecture can be modified to improve accessibility and engagement. EDLI is currently piloting such a custom GPT to explore whether it can offer effective and meaningful suggestions to instructors. Such efforts provide another example of how AI can be leveraged to empower students, rather than framing certain students are burdens to be alleviated.