September 14, 3 p.m. Ever since OpenAI released ChatGPT, the internet has been awash in synthetic text, with suggested applications including robo-lawyers, robo-therapists, and robo-journalists. University of Washington Professor Emily M. Bender posits that these applications present unacceptable risks because ChatGPT and all other language models are nothing more than ungrounded text synthesis machines. In a public lecture at the Discovery Building’s DeLuca Forum, Dr. Bender will provide an overview of how language models work, discuss risks associated with them, and explore what must hold for an appropriate use case for text synthesis. The lecture will be followed by a reception from 4-5pm.
This event is sponsored by the Language Institute, Data Science Institute, Language Sciences, and Center for the Humanities, with funding from the Anonymous Fund and Data Science Institute. For more information, visit the event website.