How to Build a Bot That Supports Good Deliberation

2min Snip

00:00

Play full episode

Summary

Transcript

Episode notes

A dese creates am a corpus for training dialogue agents so that they could understand what kind of utterances are made in these group deliberations. By annotating the uttera, as you know, being associated with asking questions, or explicitly, probing people for their reasons,. And we could then train a bot to make the kind of utterance that is associated with our successful deliberation. That's so that so blade runner sibrpont, weird, the idea that we're arguing on line. But hum, there's more efort and te newansani going to ave bot that supports good deliberation. Whywillit be a? I mani be a nice alternative to the

Link to preorder How Minds Change and get your preorder bonuses: https://www.davidmcraney.com/howmindschangehome

Deliberation. Debate. Conversation. Though it can feel like that’s what we are doing online as we trade arguments back and forth, most of the places where we currently gather make it much easier to produce arguments in isolation rather than evaluate them together in groups. The latest research suggests we will need much more of the latter if we hope to create a new, modern, functioning marketplace of ideas. In this episode, psychologist Tom Stafford takes us through his research into how to do just that.

Show notes at: http://www.youarenotsosmart.com

Patreon: http://patreon.com/youarenotsosmart