8%-10%的算法建议是 "糟糕的",但......一项探索性风险效用荟萃分析及其对监管的影响

IF 20.1 1区 管理学 Q1 INFORMATION SCIENCE & LIBRARY SCIENCE International Journal of Information Management Pub Date : 2023-12-29 DOI:10.1016/j.ijinfomgt.2023.102743
Martin Hilbert , Arti Thakur , Pablo M. Flores , Xiaoya Zhang , Jee Young Bhan , Patrick Bernhard , Feng Ji
{"title":"8%-10%的算法建议是 \"糟糕的\",但......一项探索性风险效用荟萃分析及其对监管的影响","authors":"Martin Hilbert ,&nbsp;Arti Thakur ,&nbsp;Pablo M. Flores ,&nbsp;Xiaoya Zhang ,&nbsp;Jee Young Bhan ,&nbsp;Patrick Bernhard ,&nbsp;Feng Ji","doi":"10.1016/j.ijinfomgt.2023.102743","DOIUrl":null,"url":null,"abstract":"<div><p>We conducted a quantitatively coarse-grained, but wide-ranging evaluation of the frequency recommender algorithms provide ‘good’ and ‘bad’ recommendations, with a focus on the latter. We found 151 algorithmic audits from 33 studies that report fitting risk-utility statistics from YouTube, Google Search, Twitter, Facebook, TikTok, Amazon, and others. Our findings indicate that roughly 8–10% of algorithmic recommendations are ‘bad’, while about a quarter actively protect users from self-induced harm (‘do good’). This average is remarkably consistent across the audits, irrespective of the platform nor on the kind of risk (bias/ discrimination, mental health and child harm, misinformation, or political extremism). Algorithmic audits find negative feedback loops that can ensnare users into spirals of ‘bad’ recommendations (or being ‘dragged down the rabbit hole’), but also highlight an even larger likelihood of positive spirals of ‘good recommendations’. While our analysis refrains from any judgment of the causal consequences and severity of risks, the detected levels surpass those associated with many other consumer products. They are comparable to the risk levels of generic food defects monitored by public authorities such as the FDA or FSIS in the United States. Consequently, our findings inform the ongoing discussion regarding regulatory oversight of the potential risks posed by recommender algorithms.</p></div>","PeriodicalId":48422,"journal":{"name":"International Journal of Information Management","volume":"75 ","pages":"Article 102743"},"PeriodicalIF":20.1000,"publicationDate":"2023-12-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.sciencedirect.com/science/article/pii/S026840122300124X/pdfft?md5=9e64a3c50fa338d21d87fc17bfb9b986&pid=1-s2.0-S026840122300124X-main.pdf","citationCount":"0","resultStr":"{\"title\":\"8–10% of algorithmic recommendations are ‘bad’, but… an exploratory risk-utility meta-analysis and its regulatory implications\",\"authors\":\"Martin Hilbert ,&nbsp;Arti Thakur ,&nbsp;Pablo M. Flores ,&nbsp;Xiaoya Zhang ,&nbsp;Jee Young Bhan ,&nbsp;Patrick Bernhard ,&nbsp;Feng Ji\",\"doi\":\"10.1016/j.ijinfomgt.2023.102743\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<div><p>We conducted a quantitatively coarse-grained, but wide-ranging evaluation of the frequency recommender algorithms provide ‘good’ and ‘bad’ recommendations, with a focus on the latter. We found 151 algorithmic audits from 33 studies that report fitting risk-utility statistics from YouTube, Google Search, Twitter, Facebook, TikTok, Amazon, and others. Our findings indicate that roughly 8–10% of algorithmic recommendations are ‘bad’, while about a quarter actively protect users from self-induced harm (‘do good’). This average is remarkably consistent across the audits, irrespective of the platform nor on the kind of risk (bias/ discrimination, mental health and child harm, misinformation, or political extremism). Algorithmic audits find negative feedback loops that can ensnare users into spirals of ‘bad’ recommendations (or being ‘dragged down the rabbit hole’), but also highlight an even larger likelihood of positive spirals of ‘good recommendations’. While our analysis refrains from any judgment of the causal consequences and severity of risks, the detected levels surpass those associated with many other consumer products. They are comparable to the risk levels of generic food defects monitored by public authorities such as the FDA or FSIS in the United States. Consequently, our findings inform the ongoing discussion regarding regulatory oversight of the potential risks posed by recommender algorithms.</p></div>\",\"PeriodicalId\":48422,\"journal\":{\"name\":\"International Journal of Information Management\",\"volume\":\"75 \",\"pages\":\"Article 102743\"},\"PeriodicalIF\":20.1000,\"publicationDate\":\"2023-12-29\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"https://www.sciencedirect.com/science/article/pii/S026840122300124X/pdfft?md5=9e64a3c50fa338d21d87fc17bfb9b986&pid=1-s2.0-S026840122300124X-main.pdf\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"International Journal of Information Management\",\"FirstCategoryId\":\"91\",\"ListUrlMain\":\"https://www.sciencedirect.com/science/article/pii/S026840122300124X\",\"RegionNum\":1,\"RegionCategory\":\"管理学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"INFORMATION SCIENCE & LIBRARY SCIENCE\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"International Journal of Information Management","FirstCategoryId":"91","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S026840122300124X","RegionNum":1,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"INFORMATION SCIENCE & LIBRARY SCIENCE","Score":null,"Total":0}
引用次数: 0

摘要

我们对推荐算法提供 "好 "和 "坏 "推荐的频率进行了粗粒度但广泛的定量评估,重点关注后者。我们从 33 项研究中找到了 151 项算法审核,这些研究报告了来自 YouTube、谷歌搜索、Twitter、Facebook、TikTok、亚马逊等的风险效用拟合统计数据。我们的研究结果表明,约有 8%-10% 的算法推荐是 "坏的",而约有四分之一的算法推荐积极保护用户免受自我伤害("好的")。这一平均值在所有审计中都非常一致,与平台和风险类型(偏见/歧视、心理健康和儿童伤害、错误信息或政治极端主义)无关。算法审计发现,负面反馈循环可能会使用户陷入 "坏 "推荐(或被 "拖入兔子洞")的漩涡,但同时也凸显了 "好推荐 "的正面漩涡的更大可能性。虽然我们的分析没有对风险的因果关系和严重程度做出任何判断,但检测到的风险水平超过了许多其他消费品。它们与美国 FDA 或 FSIS 等公共机构监测的普通食品缺陷的风险水平相当。因此,我们的研究结果为正在进行的有关推荐算法潜在风险监管的讨论提供了信息。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
查看原文
分享 分享
微信好友 朋友圈 QQ好友 复制链接
本刊更多论文
8–10% of algorithmic recommendations are ‘bad’, but… an exploratory risk-utility meta-analysis and its regulatory implications

We conducted a quantitatively coarse-grained, but wide-ranging evaluation of the frequency recommender algorithms provide ‘good’ and ‘bad’ recommendations, with a focus on the latter. We found 151 algorithmic audits from 33 studies that report fitting risk-utility statistics from YouTube, Google Search, Twitter, Facebook, TikTok, Amazon, and others. Our findings indicate that roughly 8–10% of algorithmic recommendations are ‘bad’, while about a quarter actively protect users from self-induced harm (‘do good’). This average is remarkably consistent across the audits, irrespective of the platform nor on the kind of risk (bias/ discrimination, mental health and child harm, misinformation, or political extremism). Algorithmic audits find negative feedback loops that can ensnare users into spirals of ‘bad’ recommendations (or being ‘dragged down the rabbit hole’), but also highlight an even larger likelihood of positive spirals of ‘good recommendations’. While our analysis refrains from any judgment of the causal consequences and severity of risks, the detected levels surpass those associated with many other consumer products. They are comparable to the risk levels of generic food defects monitored by public authorities such as the FDA or FSIS in the United States. Consequently, our findings inform the ongoing discussion regarding regulatory oversight of the potential risks posed by recommender algorithms.

求助全文
通过发布文献求助,成功后即可免费获取论文全文。 去求助
来源期刊
International Journal of Information Management
International Journal of Information Management INFORMATION SCIENCE & LIBRARY SCIENCE-
CiteScore
53.10
自引率
6.20%
发文量
111
审稿时长
24 days
期刊介绍: The International Journal of Information Management (IJIM) is a distinguished, international, and peer-reviewed journal dedicated to providing its readers with top-notch analysis and discussions within the evolving field of information management. Key features of the journal include: Comprehensive Coverage: IJIM keeps readers informed with major papers, reports, and reviews. Topical Relevance: The journal remains current and relevant through Viewpoint articles and regular features like Research Notes, Case Studies, and a Reviews section, ensuring readers are updated on contemporary issues. Focus on Quality: IJIM prioritizes high-quality papers that address contemporary issues in information management.
期刊最新文献
Editorial Board Influence of perceived risk on consumers’ intention and behavior in cross-border e-commerce transactions: A case study of the Tmall Global platform Collaborative AI in the workplace: Enhancing organizational performance through resource-based and task-technology fit perspectives Personal data strategies in digital advertising: Can first-party data outshine third-party data? Using the influence of human-as-machine representation for self-improvement products
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
现在去查看 取消
×
提示
确定
0
微信
客服QQ
Book学术公众号 扫码关注我们
反馈
×
意见反馈
请填写您的意见或建议
请填写您的手机或邮箱
已复制链接
已复制链接
快去分享给好友吧!
我知道了
×
扫码分享
扫码分享
Book学术官方微信
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术
文献互助 智能选刊 最新文献 互助须知 联系我们:info@booksci.cn
Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。
Copyright © 2023 Book学术 All rights reserved.
ghs 京公网安备 11010802042870号 京ICP备2023020795号-1