2021 IEEE International Symposium on Technology and Society (ISTAS)最新文献_第3页

Investigating accuracy disparities for gender classification using convolutional neural networks 使用卷积神经网络研究性别分类的准确性差异

2021 IEEE International Symposium on Technology and Society (ISTAS) Pub Date : 2021-10-28 DOI: 10.1109/istas52410.2021.9629153

Lia Chin-Purcell, America Chambers

{"title":"Investigating accuracy disparities for gender classification using convolutional neural networks","authors":"Lia Chin-Purcell, America Chambers","doi":"10.1109/istas52410.2021.9629153","DOIUrl":"https://doi.org/10.1109/istas52410.2021.9629153","url":null,"abstract":"Automatic gender recognition (AGR) is a subfield of facial recognition that has recently been scrutinized for bias in the form of misgendering and erasure against various identity groups in our society. Recent studies have found that several commercial AGR classifiers (from Microsoft, IMB, Face++) are biased against women and darker-skinned people as well as gender non-binary people [8, 11]. In this work, we investigate and quantify AGR classifier bias against transgender people by developing and evaluating three different convolutional neural networks (CNN): using images of cisgender individuals, using images of transgender individuals, and using images of both cisgender and transgender individuals. We find that the cisgender trained classifier is 91.7% accurate when evaluated on cisgender people, but only 68.9% accurate when evaluated on transgender people, with the worst performance of 38.6% precision for transgender men. We investigate this low precision further by performing additional experiments where various parts of the face are obscured. We end with recommendations for commercial classifiers based upon our findings.","PeriodicalId":314239,"journal":{"name":"2021 IEEE International Symposium on Technology and Society (ISTAS)","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127280670","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

From value-lists to value-based engineering with IEEE 7000™ 从价值列表到基于价值的工程与IEEE 7000™

2021 IEEE International Symposium on Technology and Society (ISTAS) Pub Date : 2021-10-28 DOI: 10.1109/istas52410.2021.9629134

S. Spiekermann

{"title":"From value-lists to value-based engineering with IEEE 7000™","authors":"S. Spiekermann","doi":"10.1109/istas52410.2021.9629134","DOIUrl":"https://doi.org/10.1109/istas52410.2021.9629134","url":null,"abstract":"Digital ethics is currently being discussed worldwide as a necessity to create more reliable IT systems. This discussion, fueled by the fear of uncontrollable general artificial intelligence (AI) and by ethical dilemmas of existing systems, has moved many institutions and scientists to demand value principles that should guide the development of future IT systems. These usually include the demand for privacy, security, transparency, fairness, etc. This article shows why working through lists of values is insufficient for good or ethically aligned design. It will be shown how a truly ethical ‘Value-based Engineering’ (VbE) would have to look like instead, so that technical product innovation as a whole is put on better (more ethical) feet. VbE is a process-driven, holistic approach to system engineering which initially drew from the ideas of Value Sensitive Design and Ethical Computing. From 2016-2021 VbE was further fleshed out in the IEEE 7000™standardization project *.*This article presents inter alia guidance for ethical engineering given in the forthcoming IEEE 7000− standard. However, this article solely represents the views of the author and does not necessarily represent a position of either the IEEE P7000 Working Group, IEEE or the IEEE Standards Association. The official link to the IEEE P7000 is: https://sagroups.iece.org/7000/.","PeriodicalId":314239,"journal":{"name":"2021 IEEE International Symposium on Technology and Society (ISTAS)","volume":"41 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131500107","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 2

Evaluating the disruption of COVID-19 on AI innovation using patent filings 利用专利申请评估COVID-19对人工智能创新的影响

2021 IEEE International Symposium on Technology and Society (ISTAS) Pub Date : 2021-10-28 DOI: 10.1109/istas52410.2021.9629125

M. Alexopoulos, Kelly A. Lyons, Kaushar Mahetaji, Keli Chiu

{"title":"Evaluating the disruption of COVID-19 on AI innovation using patent filings","authors":"M. Alexopoulos, Kelly A. Lyons, Kaushar Mahetaji, Keli Chiu","doi":"10.1109/istas52410.2021.9629125","DOIUrl":"https://doi.org/10.1109/istas52410.2021.9629125","url":null,"abstract":"Economists have long recognized that technological innovation is a key contributor to economic growth due to its impact on productivity. In this paper, we explore the impact of COVID-19 on innovation in artificial intelligence (AI) to better understand future effects on economic growth and productivity. Using patents as a measure of innovation and knowledge production, we analyze monthly patent application filing data from January 2015 to June 2021 to compare and assess trends. Past research has shown that growth in patents in the fields of AI have accelerated since 2012, with 6.5 times more annual filings occurring from 2006 to 2017. Here, we focus specifically on determining if the pandemic has had an impact on this acceleration in AI-related innovation. To accomplish this task we must confront the challenge in using up-to-date patent data for this kind of analysis due to the fact that there are considerable time lags associated with patent filing dates and their ultimate publication dates. In real-time situations such as COVID-19, it is, therefore, difficult to ascertain impact using the publicly available patenting data directly. In this paper, we propose a novel approach for examining existing and up-to-date publicly available patent filing data and use that method to gain new insights into the pandemic’s effects on AI-related innovation. Our findings suggest that the pandemic has had a slowing impact on the rate of innovation in these areas but that the downturn may be reversing.","PeriodicalId":314239,"journal":{"name":"2021 IEEE International Symposium on Technology and Society (ISTAS)","volume":"36 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132412650","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 1

Ethics of AI as practical ethics 人工智能伦理作为实践伦理

2021 IEEE International Symposium on Technology and Society (ISTAS) Pub Date : 2021-10-28 DOI: 10.1109/istas52410.2021.9629163

Florian Richter

{"title":"Ethics of AI as practical ethics","authors":"Florian Richter","doi":"10.1109/istas52410.2021.9629163","DOIUrl":"https://doi.org/10.1109/istas52410.2021.9629163","url":null,"abstract":"Whereas we acknowledge in general certain values as crucial for the ethical debate in AI, like e.g., fairness, transparency, and accountability, these values often can stand in conflict with each other. E.g., more transparency can lead to less privacy. Introducing higher principles to balance the values faces two problematic issues: (1.) Principles also can stand in conflict with each other and defer the discussion into a purely theoretical realm. (2.) If a higher-level principle is introduced and this also stands in conflict with another principle, then a higher-higher-level principle is needed, and we will get into an infinite regress. Although ethics of AI is part of the so-called field of applied ethics and it seems therefore that it is about the application of principles and values and finding the right balance regarding certain ethical theories, e.g., Kantian ethics or utilitarianism ([1][2][3]), the traditional approaches in the field of applied ethics do not offer sufficient conceptual means to deal with practical problems. Thus, the problems that arise from the implementation of intelligent systems can also not be handled adequately, because the same issue as above arises: How can the values be balanced regarding the ethical theories? If higher-level principles are not a viable approach to resolve conflicts of values, the criteria under which they can be implemented should be taken into consideration. Therefore, it is proposed that specific criteria for the implementation of the values need to be made explicit, which will result at least in a clarification for the public debate about certain technological advancements in the field of AI. [4] Furthermore, it is proposed that to resolve these conflicts the implementation must be evaluated, whether it enables human intervention in the future instead of making further actions and interventions impossible.","PeriodicalId":314239,"journal":{"name":"2021 IEEE International Symposium on Technology and Society (ISTAS)","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114704019","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 1

A review of data governance challenges in smart farming and potential solutions 回顾智能农业中的数据治理挑战和潜在解决方案

2021 IEEE International Symposium on Technology and Society (ISTAS) Pub Date : 2021-10-28 DOI: 10.1109/istas52410.2021.9629169

Adesola Anidu, Rozita Dara

引用次数: 4

From historical thinking to critical thinking about technology 从历史思维到对技术的批判性思维

2021 IEEE International Symposium on Technology and Society (ISTAS) Pub Date : 2021-10-28 DOI: 10.1109/istas52410.2021.9629144

S. Campbell

引用次数: 0

Fairness in AI applications 人工智能应用中的公平性

2021 IEEE International Symposium on Technology and Society (ISTAS) Pub Date : 2021-10-28 DOI: 10.1109/istas52410.2021.9629140

C. Shelley

引用次数: 0

See something, say something? Coordinating the disclosure of security vulnerabilities in Canada’s infrastructure 看到什么，说什么?协调加拿大基础设施安全漏洞的披露

2021 IEEE International Symposium on Technology and Society (ISTAS) Pub Date : 2021-10-28 DOI: 10.1109/istas52410.2021.9629214

Yuan Stevens, S. Tran, Ryan Atkinson

引用次数: 0

AI can turn the clock back before we know it 人工智能可以在我们意识到之前让时光倒流

2021 IEEE International Symposium on Technology and Society (ISTAS) Pub Date : 2021-10-28 DOI: 10.1109/istas52410.2021.9629161

A. Gerdes

引用次数: 1

Cybersecurity issues in citizen science 公民科学中的网络安全问题

2021 IEEE International Symposium on Technology and Society (ISTAS) Pub Date : 2021-10-28 DOI: 10.1109/istas52410.2021.9629198

D. Schaeffer, P. Olson

{"title":"Cybersecurity issues in citizen science","authors":"D. Schaeffer, P. Olson","doi":"10.1109/istas52410.2021.9629198","DOIUrl":"https://doi.org/10.1109/istas52410.2021.9629198","url":null,"abstract":"As Citizen Science projects become more wide-spread and global, we must pay attention to cybersecurity issues that can emerge. Sharing data is fundamental to Citizen Science projects, and that requires attention to protecting privacy. This issue is exacerbated if participants use common technologies, e.g., cell phones, that are used for other purposes and may be subject to cybersecurity vulnerabilities. Cybersecurity issues can arise during recruitment, by which social media is a prevalent tool, to Citizen Science projects in that bad actors can solicit input into malicious and false projects. Furthermore, the results of Citizen Science projects may provide foundations for policy and legislation; thus, integrity must be maintained, and bad actors prevented from manipulating data and/or results. Governance of Citizen Science projects must include attention to cybersecurity issues, regardless of their scope and scale. The exploitation of cybersecurity vulnerabilities is often the result of ethical lapses. Just as scientists respect science research ethics, so must citizens who undertake scientific research. In this paper, we will share a taxonomy that identifies how research ethics are represented and brought to life in 21st century Citizen Science projects.","PeriodicalId":314239,"journal":{"name":"2021 IEEE International Symposium on Technology and Society (ISTAS)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131066556","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0