Red Teaming and the Ethics of AI Decision-Making

First up on the podcast, producer Meagan Cantwell and Contributing Correspondent Sara Reardon discuss alternative approaches to animal testing, from a heart on a chip to a miniorgan in a dish.

Next on the show, Expert Voices columnist Melanie Mitchell and host Sarah Crespi dig into AI lies. Why do chatbots fabricate answers and pretend to do math? Mitchell describes the stress tests large language models undergo—called red teaming—and the steps needed to better understand how they “think.”  

Melanie Mitchell is a professor at the Santa Fe Institute. You can read all her Expert Voices columns here.  

This week’s episode was produced with help from Podigy.

About the Science Podcast

Authors: Sarah Crespi; Meagan Cantwell; Melanie Mitchell; Sara Reardon

Learn more about your ad choices. Visit megaphone.fm/adchoices