Salvato in:
Dettagli Bibliografici
Autori principali: Cho, Hyundong, Liu, Shuai, Shi, Taiwei, Jain, Darpan, Rizk, Basem, Huang, Yuyang, Lu, Zixun, Wen, Nuan, Gratch, Jonathan, Ferrara, Emilio, May, Jonathan
Natura: Preprint
Pubblicazione: 2023
Soggetti:
Accesso online:https://arxiv.org/abs/2311.10781
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866916235912937472
author Cho, Hyundong
Liu, Shuai
Shi, Taiwei
Jain, Darpan
Rizk, Basem
Huang, Yuyang
Lu, Zixun
Wen, Nuan
Gratch, Jonathan
Ferrara, Emilio
May, Jonathan
author_facet Cho, Hyundong
Liu, Shuai
Shi, Taiwei
Jain, Darpan
Rizk, Basem
Huang, Yuyang
Lu, Zixun
Wen, Nuan
Gratch, Jonathan
Ferrara, Emilio
May, Jonathan
contents Conversational moderation of online communities is crucial to maintaining civility for a constructive environment, but it is challenging to scale and harmful to moderators. The inclusion of sophisticated natural language generation modules as a force multiplier to aid human moderators is a tantalizing prospect, but adequate evaluation approaches have so far been elusive. In this paper, we establish a systematic definition of conversational moderation effectiveness grounded on moderation literature and establish design criteria for conducting realistic yet safe evaluation. We then propose a comprehensive evaluation framework to assess models' moderation capabilities independently of human intervention. With our framework, we conduct the first known study of language models as conversational moderators, finding that appropriately prompted models that incorporate insights from social science can provide specific and fair feedback on toxic behavior but struggle to influence users to increase their levels of respect and cooperation.
format Preprint
id arxiv_https___arxiv_org_abs_2311_10781
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Can Language Model Moderators Improve the Health of Online Discourse?
Cho, Hyundong
Liu, Shuai
Shi, Taiwei
Jain, Darpan
Rizk, Basem
Huang, Yuyang
Lu, Zixun
Wen, Nuan
Gratch, Jonathan
Ferrara, Emilio
May, Jonathan
Computation and Language
Artificial Intelligence
Conversational moderation of online communities is crucial to maintaining civility for a constructive environment, but it is challenging to scale and harmful to moderators. The inclusion of sophisticated natural language generation modules as a force multiplier to aid human moderators is a tantalizing prospect, but adequate evaluation approaches have so far been elusive. In this paper, we establish a systematic definition of conversational moderation effectiveness grounded on moderation literature and establish design criteria for conducting realistic yet safe evaluation. We then propose a comprehensive evaluation framework to assess models' moderation capabilities independently of human intervention. With our framework, we conduct the first known study of language models as conversational moderators, finding that appropriately prompted models that incorporate insights from social science can provide specific and fair feedback on toxic behavior but struggle to influence users to increase their levels of respect and cooperation.
title Can Language Model Moderators Improve the Health of Online Discourse?
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2311.10781