22: Cloaking Your Photos with Fawkes - Language Models

In the first part of the show, we talk to Shawn Shan, a PhD student at University of Chicago, who works in privacy and security. Shawn is one of the researchers behind Fawkes, which is software that makes subtle changes to images to make it harder for machines to recognize images as the same, even if to the human eye they appear identical. 

In the second half of the show, Dr. Peter Eckersley is back to talk about Language Models. Back in July, Open AI, the AI research and deployment company based in San Francisco, released the 3rd generation of their language model, GPT-3. This version has garnered a lot of excitement. There’s been people who have used it to generate conversations between computing pioneers Claude Shannon and Alan Turing or having it generate entire stories with just a one sentence prompt.