Ep. 39: Can We Trust AI Judgments?


Manage episode 243386980 series 1260429
By Atief Heermance, Robert de Neufville, Scott Eastman, Atief Heermance, Robert de Neufville, and Scott Eastman. Discovered by Player FM and our community — copyright is owned by the publisher, not Player FM, and audio is streamed directly from their servers. Hit the Subscribe button to track updates in Player FM, or paste the feed URL into other podcast apps.
Episode 39 of the NonProphets podcast, in which Atief, Robert, and Scott announce that IARPA's new Hybrid Forecasting Competition is looking for volunteer forecasters (00:19), interrupt their own discussion to explain that some of the podcast audio was lost due to what they believe was an attack by a rogue AI (06:28), discuss to what extent we can trust the output of AI forecasters (08:09), talk about the "value-alignment problem" in AI design (21:39), and return to the question of how much we can trust AI before the podcast is mysteriously cut off (26:46). As always, you can reach us at nonprophetspod.wordpress.com, or nonprophetspod@gmail.com. (recorded 8/16/2017)

162 episodes