Part 3 of the TED Radio Hour episode Cuts Both Ways.
If AI is smart enough to pass the bar exam, why does it struggle with simple common sense questions? Computer scientist Yejin Choi studies how to teach AI human reasoning, social norms and values.
About Yejin Choi
Yejin Choi is a computer science professor at the University of Washington. She is also a MacArthur Fellow, a distinguished research fellow at the Institute for Ethics in AI at the University of Oxford, and a senior research director at the nonprofit Allen Institute for AI, where she oversees The Mosaic Project. Choi is the co-recipient of two Test of Time Awards and five best paper awards from top academic conferences in AI. She won the inaugural Alexa Prize Challenge in 2017, was a co-recipient of the Borg Early Career Award in 2018 and was named to the Institute of Electrical and Electronics Engineers' "10 to Watch" in AI list in 2016.
Note: For this segment, host Manoush Zomorodi had a conversation with GPT-4, the latest update in the large-language models that make up Open AI's Chat GPT interface.
This segment of TED Radio Hour was produced by Harsha Nahata and edited by Sanaz Meshkinpour and James Delahoussaye. You can follow us on Facebook @TEDRadioHour and email us at TEDRadioHour@npr.org.