当前位置:首页 > 报告详情

AI的安全与隐私问题考量.pdf

上传人: a****d 编号:184996 2024-10-07 14页 524.03KB

1、1|2024 SNIA.All Rights Reserved.Security and Privacy Concerns for AI Eric Hibbard,CISSP,FIP,CISASamsung Semiconductor,Inc.2|2024 SNIA.All Rights Reserved.IntroductionArtificial intelligence(AI)systems are creating numerous opportunities and challenges for many facets of society.For security,AI is pr

2、oving to be a power tool for both adversaries and defenders.Privacy is similar,but the societal concerns are elevated to a point where laws and regulations are already being enacted.3|2024 SNIA.All Rights Reserved.AI Ethical and Societal Concerns Ethical and societal concerns are a factor when devel

3、oping and using AI systems and applications Taking context,scope and risks into consideration can mitigate undesirable ethical and societal outcomes and harms such as:financial harm psychological harm harm to physical health or safety intangible property(for example,IP theft,damage to a companys rep

4、utation)social or political systems(for example,election interference,loss of trust in authorities)civil liberties(for example,unjustified imprisonment or other punishment,censorship,privacy breaches)Source:ISO/IEC TR 24368:2022 Information technology Artificial Intelligence Overview of ethical and

5、societal concerns4|2024 SNIA.All Rights Reserved.Examples of potential harms related to AI systemsSource:NIST Artificial Intelligence Risk Management Framework(AI RMF 1.0)5|2024 SNIA.All Rights Reserved.OECD and Artificial IntelligenceThe OECD AI PrinciplesOECD Framework for the Classification of AI

6、 SystemsSource:OECD(2022)OECD Framework for the Classification of AI systems OECD Digital Economy PapersThe OECD has developed a framework for classifying AI lifecycle activities according to five key socio-technical dimensions,each with properties relevant for AI policy and governance,including ris

word格式文档无特别注明外均可编辑修改,预览文件经过压缩,下载原文更清晰!
三个皮匠报告文库所有资源均是客户上传分享,仅供网友学习交流,未经上传用户书面授权,请勿作商用。
本文主要探讨了人工智能(AI)系统在安全、隐私以及伦理和社会方面的挑战和风险。AI系统在为社会带来机遇的同时,也给安全领域带来了强大的挑战,包括数据隐私泄露、社会和政治系统的干扰等问题。在伦理和社会层面,文章强调了在AI系统的开发和使用过程中,考虑其上下文、范围和风险是减轻不良后果的关键。这包括避免金融、心理伤害,保护物理安全,维护知识产权,以及保护公民自由等。 文章还提到了OECD和NIST对AI风险管理的框架,以及ISO/IEC对AI生命周期和概念的定义。AI风险与传统软件风险不同,主要体现在数据代表性、系统依赖性、训练过程中的变化、数据集的时效性以及系统的规模和复杂性等方面。AI系统的攻击类型包括污染攻击、逃避攻击、成员推断、模型外泄和模型反转等。 在隐私方面,AI系统面临的威胁包括可识别性、链接性、非否认性、可检测性、信息披露、不知情处理以及不合规等。文章指出,AI系统的可信度要求是有效和可靠,同时需要可追溯和透明。 目前,AI的发展趋势和立法动向表明,政府和监管机构正在关注AI的安全性和伦理性,如欧盟的AI Act、美国的行政命令,以及多个州关于AI的立法。此外,知识产权问题和自主AI系统的兴起也是当前AI领域的热点。
"AI伦理与社会担忧有哪些?" "AI系统面临哪些安全风险?" "AI发展趋势与政策法规如何?"
客服
商务合作
小程序
服务号
折叠