Katy Gero

Research Abstract:

Writing is a deeply human activity centered on communicative intent: we strive to make ourselves understood by others. This striving is about our desire to be heard and seen. While computers can generate fluent and convincing text, they lack a reason to write. My research draws on a cognitive process model of writing to help people perform complex and high-impact writing tasks with computational support, as well as design ethical technologies that support writing in ways that writers feel are authentic and meaningful. My research focuses on questions that combine technological innovation with psychological models of writing as a cognitive process: 1) How can we design novel interactive systems that support complex writing tasks, like generating new metaphorical explanations or explaining technical topics to a general audience? 2) How can we address the sociotechnical issues of writing support and consider writer agency, authenticity, and (mis)understanding of system capabilities?

Bio:

Katy Gero studies human-centered AI, with a focus on technology for impactful writing and understanding the limits and capabilities of large language models. She is currently a Postdoctoral Fellow in Computer Science with Elena Glassman at Harvard University. She recently completed her PhD at Columbia University with Lydia Chilton. She has been supported by an NSF Graduate Research Fellowship, the Brown Institute for Media Innovation, and Amazon Science.