Content moderation as language policy

Author:

Mandy Lau

Abstract

Commercial content moderation removes harassment, abuse, hate, or any material deemed harmful or offensive from user-generated content platforms. A platform’s content policy and related government regulations are forms of explicit language policy. This kind of policy dictates the classifications of harmful language and aims to change users’ language practices by force. However, the de facto language policy is the actual practice of language moderation by algorithms and humans. Algorithms and human moderators enforce which words (and thereby, content) can be shared, revealing the normative values of hateful, offensive, or free speech and shaping how users adapt and create new language practices. This paper will introduce the process and challenges of commercial content moderation, as well as Canada’s proposed Bill C-36 with its complementary regulatory framework, and briefly discuss the implications for language practices.

Publisher

York University Libraries

Reference62 articles.

1. Abid, A., Farooqi, M., & Zou, J. (2021). Large language models associate Muslims with violence. Nature Machine Intelligence, 3(6), 461–463. https://doi.org/10.1038/s42256-021-00359-2

2. Andrews, L. (2020, August 31). Paedophiles are using cheese and pizza emojis to communicate secretly on Instagram. Daily Mail Online. https://www.dailymail.co.uk/news/article-8681535/Paedophiles-using-cheese-pizza-emojis-communicate-secretly-Instagram.html

3. Bayer, J., & Bárd, P. (2020). Hate speech and hate crime in the EU and the evaluation of online content regulation approaches (PE655.135 - July 2020). European Parliament’s Policy Department for Citizens’ Rights and Constitutional Affairs. https://www.europarl.europa.eu/RegData/etudes/STUD/2020/655135/IPOL_STU(2020)655135_EN.pdf

4. Bender, E., Gebru, T., McMillan-Major, A., Shmitchell, S., & Anonymous. (2020). On the dangers of stochastic parrots: Can language models be too big? Proceedings of the ACM/IEEE Joint Conference on Digital Libraries, 1(1), 271–278. https://doi.org/10.1145/3442188.3445922

5. Benjamin, R. (2019). Race after technology: Abolitionist tools for the new Jim code. Polity.

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3