总体国家安全观视域下的算法安全与治理

随着算法的不断演化及其在计算、通信、仿真、数据分析、信息管理、信号处理及指挥控制等领域的不断优化,有关算法所引发的一系列安全问题也开始引起人们的关注。算法安全源自于人工智能相关技术在社会生活各个领域的深度嵌入所带来的安全问题,从国家安全视角来看,算法安全具有逻辑操纵的政治性、行为影响的颠覆性、场域渗透的隐蔽性等显著特点,并且会深度影响国家安全的各个维度。具体而言,算法安全对国家安全的挑战表现在深刻改变国家间权力结构对比、导致隐性算法依赖和逆向算法攻击等方面。近年来,随着算法安全风险逐渐显现,加强算法安全治理已经成为国际社会的共识。在结合算法安全的特点和世界各国有关算法安全的应对举措的基础上,我国可以通过提高算法风险意识、培养算法专业人才、完善算法监督体系等方式来进行有关算法安全的秩序重构,以强化人工智能时代的国家安全治理。     

论网络商业方法的可专利性

随着Internet的发展 ,电子商务成为人们关注的焦点 ,而与网络有关的商业方法专利 ,也成为热门话题。然而 ,尽管商业方法已成为美国法律的法定保护客体 ,对于这种新兴专利合法性的讨论 ,仍是方兴未艾。     

私法制度中的代数算法黑箱及其应对

因私法制度中代数算法规范(代数规范)的缺陷,一些指导性案例未能给出判决的明确算法,私法制度中的代数算法黑箱导致了判决书中的算法瑕疵。传统私法制度深受古希腊几何学的影响,因而疏离了代数算法且轻视相应的代数规范,人工智能中机器算法的繁荣无助于缓解私法中的代数算法黑箱,反而将新的算法黑箱叠加于传统黑箱之上。代数规范是私法制度中规定定量计算的法律规范,它是私法规范中的定量维度和必不可少的组成部分,且随着代数算法在认识论中的地位上升而日益重要。私法制度应当改变疏离代数算法的传统,完善其代数规范,适用代数规范作为判决的依据,并结合定性规范以指导与检验法律人工智能,以限缩判决书中的算法瑕疵。     

立法伦理与算法正义——算法主体行为的法律规制

人类设计算法、运用算法的最终目的是为了增强自身解决问题的能力,进一步提升生活水平并优化社会结构.算法黑箱存在的前提预设是人类对因果关系的认知断裂,算法黑箱针对自动化决策过程中特定的算法用户.在一系列的数据代码背后是算法主体的行为目标和价值选择,不同价值选择加上算法的涌现性和自主性的复杂特质导致相关法律关系主体的权利冲突...     

算法歧视的司法审查——意大利户户送有限责任公司算法歧视案评析

意大利户户送有限责任公司算法歧视案所涉及的众多理论问题中,以户户送有限责任公司平台使用的算法是否对骑手造成歧视为核心。博洛尼亚法院的判决从骑手是否受反歧视法律的保护、歧视的含义以及平台算法对骑手的影响是否构成歧视等角度层层深入说理论证,最终认定该平台使用的算法构成间接歧视,并要求户户送有限责任公司支付原告三个工会组织5万欧元的惩罚性赔偿。该判决是欧洲针对算法歧视做出的第一个判决,在认定算法歧视、保护骑手权利以及确认惩罚性损害赔偿方面具有重要意义。     

智慧司法中的忧虑:想象、剖析与展望

智慧司法建设如火如荼的同时,学术界对智慧司法可能带来的负面效应过分忧虑:担忧对诉讼参与人贴标签会出现主观偏见、担心裁判算法黑箱可能会产生歧视、忧心人工智能代替法官思考甚至取代法官。上述忧虑乃是基于对智慧司法的想象。事实上,贴标签并不会产生偏见;智慧司法确实充满了黑箱,但并不构成逻辑困扰;智慧司法并没有增加裁判过程中的歧视,而是限制了本就存在的歧视行为;算法没有能力代替法官思考,更不可能取代法官。未来在讨论智慧司法建设时应当破除对新鲜事物的恐惧,摘除针对算法的有色眼镜,关注算法对司法带来的各种冲击。     

The future decisions of RoboJudge HHJ Arthur Ian Blockchain: Dread,delight or derision?,

Steve Saxby's prescient founding of CLSR, two hundred issues ago, encouraged and resonated with my own digital visionary thinking and professional activity in the evolving field of ICT and the Law. From Infolex, the UK's first commercially-available computer-assisted legal information retrieval service, and my APPEAL Report (on the admissibility of computer evidence in court and the legal reliability/security of IT systems), via my Forensic Systems Analysis expert methodology, to the nascent CryptoBlockTV, Steve's scholarly foresight in promoting adventurous exploration of ‘digilaw’ high-ground topics and issues has presented me with opportunities to generate a stream of prescient material, for which I am immensely grateful. And what is beyond prescient today is that the Coming of the Robots is unstoppable. The Artificial Intelligence (AI) Age is upon us; RoboJudge has all but already arrived. While many are concerned about defining and developing Machine Ethics, Castell's Second Dictum: “You cannot construct an algorithm that will reliably decide whether or not any algorithm is ethical” reveals that this is a futile exercise. Algorithms are also pivotal to the current mania for Crypto-Algorithmic Blockchain Technology Initial Coin Offerings (ICOs), with a ‘Crypto Tribe’ of Millennials relentlessly raising billions in real money thereby, to the extent that I have dubbed Crypto the Millennials’ Rock'n’Roll. The seasoned ICT expert professional however bears in mind that there are as yet no ISO standards for blockchain, and there is far more to creating and delivering a complete quality-assured system than just the blockchain component. Furthermore, the legal status of cryptocurrency, smart contract and distributed ledger technology is not clear or uncontentious – and there is already ICO litigation on foot. Nevertheless, taking my limerick-writing Castell GhostWriteBot’s advice, it is perhaps time for my own asset-linked ICO, to launch my CapChere.com concept designed to reboot Capitalism and achieve ubiquitous universal share and wealth ownership. Look out for Castell GhostWriteBot’s account (with or without limericks) of how I fared, in the 400th issue of CLSR.     

决策树算法在未成年人犯罪成因中的应用研究

犯罪成因的分析在未成年人犯罪研究中起着重要的作用，而分类算法在分析犯罪成因上是卓有成效的。决策树算法是分类算法中一类重要的算法，其中以C4．5算法最为流行，但这种算法在时间复杂性和增量式学习方面还不是很理想，因此有必要对其进行改进。     

大数据技术在刑事定罪中的运用

近年来,在金融创新特别是互联网金融创新大旗的指引下,网络现货投资交易平台迅速增长,因缺乏有效金融监管,现货投资交易平台引发的违法犯罪数量也随之暴增。由于网络现货投资交易行为的复杂性和专业性,司法机关查处网络现货投资交易平台中的犯罪时,在罪与非罪、此罪与彼罪判定方面分歧较大,传统的定罪方法和标准已经显得力不从心。大数据技术的出现为此类复杂犯罪定罪提供了方法与标准。大数据技术不仅能提高定罪准确性,也能提高定罪效率。针对当前出现的诸如网络虚拟投资交易案等网络犯罪特征,运用数据挖掘、决策树CHAID算法等数据技术对新型网络犯罪判决进行定量分析,我们可以将大数据技术作为司法定罪裁判的辅助工具,加以推广运用。     

From Responsible Disclosure Policy (RDP) towards State Regulated Responsible Vulnerability Disclosure Procedure (hereinafter – RVDP): The Latvian approach

Cybersecurity is an integral part of security. It plays a tremendous role in modern society. It encompasses technical, organizational and legislative measures created for the purpose of protecting and minimizing impacts from cyber incidents. Any software may contain bugs or security holes. Hackers frequently discover such flaws and, without vendor's consent, disclose step-by-step instructions about vulnerability to the public, disregarding the possible IT security risk. Many vendors already have introduced responsible disclosure policies or “bug bounty” programs. In 2013 the Netherlands launched the first state responsible disclosure Guidelines. Guidelines contain principles, definitions and organizational measures, necessary for responsible disclosure policy as a state policy.Latvia decided to draft Regulation on responsible disclosure procedure. In March 2016, the Ministry of Defence created a working group. The goal of the drafters was: 1) to prepare amendment to Law on the Security of Information Systems to create legislative framework for responsible vulnerability disclosure process; 2) to draft an amendment to Section 241 (3) of Criminal Law to create a guaranty against prosecution (waiver) for persons who act in accordance with responsible disclosure process. The paper provides an insight into this process, difficulties faced by drafters and presents provisional results of the legislative draft and lessons to be learnt.