R1: Leveraging and Personalizing Generative AI for Accessibility
Background
Researcher Dr. Hernisa Kacorri is currently working on projects related to Artificial Intelligence (AI) and people with disabilities. AI (specifically generative AI) is growing in popularity. More organizations are incorporating AI models into their systems. Generative AI provides responses based on data that was inputted into it. What makes it different from doing a Google search is that it can be trained to adjust and learn from responses and new data. However, since AI models are developed by humans, they are bound to have faults and fallacies. Many of the modern AI models are designed by able-bodied individuals, sometimes resulting in bias towards the disabled community. Ableism is systematically introduced from conception into these technologies as they focus on representing or catering to the majority of technology users, marginalizing people with disabilities. Furthermore, there is a lack of research around personalization of AI, which ideally would allow people with varying abilities to adjust the applications to suit their needs.
Current Work
Work on this project will include having conversations with experts in artificial intelligence who are currently involved in creating technologies for people with disabilities. Exploration of not only how AI is biased but also how AI is adopted and distributed will help shape future accessibility. Researchers will also work directly with people with disabilities. Dr. Kacorri is especially interested in the Blind community, and how applications can be individualized to help them reach their goals. With the advances in AI, models and systems need to be updated to support people with disabilities. This work will help researchers discover the best way(s) to do this.
