Affordances and Risks of ChatGPT to Autistic Users (arxiv.org)

A recent study examined the implications of ChatGPT for autistic users, revealing both beneficial and risky aspects of its use. Through an analysis of 3,984 social media posts from self-identified autistic individuals, researchers applied the Technology Affordance framework to uncover how these users leverage ChatGPT as a tool for offloading executive dysfunction, regulating emotions, and navigating neurotypical communication. However, the study also identified significant risks, including the potential for technology to reinforce delusional thinking, mask authentic identity, and conflict with personal ethical values. This research is significant for the AI and machine learning community as it highlights the dual nature of large language models in supporting marginalized groups, showcasing the need for more inclusive technology design. The findings call attention to the vital balance between harnessing the utility of AI tools and addressing their risks, paving the way for the development of neuro-inclusive technologies that enhance accessibility while minimizing unintended consequences. Future work is anticipated to further explore these trade-offs and foster more supportive interaction modalities for autistic users.