How Neurodivergent Students Conceptualize the Risks and Benefits of AI Chatbots

There are many risks and benefits to using AI chatbot models. Previous research shows that neurodivergent users benefit from chatbots as a tool to serve their specific needs and provide answers in a user-friendly format. They are also at increased risk for the negative consequences of AI chatbots such as chatbot bias or overreliance on AI tools. To understand these risks and benefits, I conducted a workshop for neurodivergent college students to understand their lived experiences with AI chatbots and what future AI chatbots could do to better serve the needs of neurodivergent users. Participants reported that they use chatbots as a starting point for tasks, as a planning tool, or to break down large amounts of information. Participants observed that using chatbots to enhance the learning process is beneficial but using them to automate work is overreliance. Participants also reported that chatbots are helpful as search engines to get started on tasks such as finding social venues but relying on them as a social support risks isolation. With participants, I created five recommendations on how AI chatbots can better serve neurodivergent users which include (1) promoting learning over automation, (2) adding value to social connections without being a social connect itself, (3) serving as a starting point for finding mental health resources, (4) communicating directly to users what privacy settings are in place, and (5) offering guidance on social nuances and aid in executive functioning.