Abstract: Rapid improvements of generative AI techniques may help us unlock a more accessible world. For example, recent vision to language models can provide high quality visual descriptions for blind and low vision users, and large language models can be used to simplify text. The benefits of large pretrained generative AI tools for accessibility come with the risk of misleading users and amplifying existing societal biases. In this talk, Dr. Pavel will discuss three projects in which they explore using generative AI to make creativity (GenAssist), media consumption (ShortScribe), and real time communication (COMPA) more accessible. Dr. Pavel will share benefits that they observed that go beyond what was possible before with either manual or automated work, promising interaction techniques for mitigating the risks of using generative AI, and outstanding challenges with using generative AI that call for future work.

Bio: Amy Pavel is an Assistant Professor in the Department of Computer Science at UT Austin. Her research in Human-AI Interaction and Accessibility aims to make communication more effective and accessible. Before starting at UT, she received her PhD in CS from UC Berkeley, then worked as a Research Scientist at Apple and a Postdoctoral Fellow at Carnegie Mellon University. Her work has been recognized by Best Paper and Honorable Mention awards at ACM CHI and UIST.