The Problem With GPT Is It Returning False Information

This week on Rational Security, Alan, Quinta, and Scott were joined once again by host emeritus Benjamin Wittes to talk through the week's various freak-outs, including:

• “We Found the 100th Luftballon.” Last week, a Chinese spy balloon floated over the United States, triggering a national freak-out that led to the cancellation of a major high-level summit between U.S. and Chinese leaders. Was this freak-out warranted? What does it tell us about U.S.-China relations?
• “SotFU.” President Biden delivered his second State of the Union address last night—and it was about as contentious as expected. How did he do? And how should we feel about this most vaunted of national institutions?
• “ChatOMG.” Over the past several weeks, countless Americans have had the chance to hash it out with ChatGPT, a large language-model artificial intelligence that is open to the public and will either revolutionize or devastate a thousand different human tasks, depending on who you ask. Just how revolutionary is ChatGPT? And is that a good thing or a bad thing?

Support this show http://supporter.acast.com/lawfare.

Hosted on Acast. See acast.com/privacy for more information.