GPT-3 and InstructGPT: technological dystopianism, utopianism, and “Contextual” perspectives in AI ethics and industry

Author:

Chan AnastasiaORCID

Abstract

AbstractThis paper examines the ethical solutions raised in response to OpenAI’s language model Generative Pre-trained Transformer-3 (GPT-3) a year and a half from its release. I argue that hype and fear about GPT-3, even within the Natural Language Processing (NLP) industry and AI ethics, have often been underpinned by technologically deterministic perspectives. These perspectives emphasise the autonomy of the language model rather than the autonomy of human actors in AI systems. I highlight the existence of deterministic perspectives in the current AI discourse (which range from technological utopianism to dystopianism), with a specific focus on the two issues of: (1) GPT-3’s potential intentional misuse for manipulation and (2) unintentional harm caused by bias. In response, I find that a contextual approach to GPT-3, which is centred upon wider ecologies of societal harm and benefit, human autonomy, and human values, illuminates practical solutions to concerns about manipulation and bias. Additionally, although OpenAI’s newest 2022 language model InstructGPT represents a small step in reducing toxic language and aligning GPT-3 with user intent, it does not provide any compelling solutions to manipulation or bias. Therefore, I argue that solutions to address these issues must focus on organisational settings as a precondition for ethical decision-making in AI, and high-quality curated datasets as a precondition for less harmful language model outputs.

Funder

Macquarie University

Publisher

Springer Science and Business Media LLC

Subject

General Earth and Planetary Sciences

Cited by 46 articles. 订阅此论文施引文献 订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献

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