信息安全与通信保密

参考文献

1 See Aaron M.Kessler,Law Left Behind as HandsFree Cars Cruise,STAR TRIBUNE(May 3,2015,12:21 PM),http://www.startribune.com/law-leftbehind-as-hands-free-carscruise/302322781/[https://perma.cc/39PB-UDJ8].

2 See,e.g.,John Markoff,Armies of Expensive Lawyers,Replaced by Cheaper Software,N.Y.TIMES(Mar.4,2011),http://www.nytimes.com/2011/03/05/science/05legal.html(May 4,2016);Timothy Williams,Facial Recognition Software Moves from Overseas Wars to Local Police,N.Y.TIMES(Aug.12,2015),http://www.nytimes.com/2015/08/13/us/facialrecognitionsoftware-moves-from-overseas-wars-to-localpolice.html(May 4,2016).

3 See,e.g.,Yves Eudes,The Journalists Who Never Sleep,GUARDIAN(Sept.12,2014,6:17 AM),http://www.theguardian.com/technology/2014/sep/12/artificial-intelligence-datajournalismmedia[https://perma.cc/CES7-X58C](discussing the increasing use of“robot writers”in journalism).

4 Artificial Intelligence:Rise of the Machines,ECONOMIST(May 9,2015),http://www.economist.com/news/briefing/21650526-artificial-intelligence-scarespeopleexcessivelyso-rise-machines[https://perma.cc/B2LD-B4XS].

5 See,e.g.,John Frank Weaver,We Need to Pass Legislation on Artificial Intelligence Early and Often,SLATE(Sept.12,2014,3:53 PM),http://www.slate.com/blogs/future_tense/2014/09/12/we_need_to_pass_artificial_intelligence_laws_early_and_often.html[https://perma.cc/6SKM-K6RT];Perri 6,Ethics,Regulation and the New Artificial Intelligence,Part I:Accountability and Power,4 INFO.,COMM.&SOC’Y 199,203(2010).

6 Compare,e.g.,STUART J.RUSSELL&PETERNORVIG,ARTIFICIAL INTELLIGENCE:AMODERN APPROACH 1034(3d ed.2010)(discussing how people losing their jobs to automation is an ethical issue introduced by AI),with JOHN MAYNARD KEYNES,Economic Possibilities for Our Grandchildren,in ESSAYSON PERSUASION 321,325(1972)(“For the moment the very rapidity of these changes is hurting us and bringing difficult problems to solve...namely,technological unemployment.This means unemployment due to our discovery of means of economising the use of labour outrunning the pace at which we can find new uses for labour.”).

7 See RUSSELL&NORVIG,supra note 7,at 1035(discussion titled“AI systems might be used toward undesirable ends”).

8 Aileen Graef,Elon Musk:We Are“Summoning a Demon”with Artificial Intelligence,UPI(Oct.27,2014,7:50 AM),http://www.upi.com/Business_News/2014/10/27/Elon Musk-Weare-summoning-a-demon-with-artificialintelligence/4191414407652/[https://perma.cc/M98J-VYNH].

9 See,e.g.,Eric Mack,Bill Gates Says You Should Worry About Artificial Intelligence,FORBES(Jan.28,2015),http://www.forbes.com/sites/ericmack/2015/01/28/bill-gates-alsoworriesartificial-intelligence-is-a-threat/(quoting Bill Gates,“I am in the camp that isconcerned....First the machines will do a lot of jobs for us and not be super intelligent.That should be positive if we manage it well.A few decades after that though the intelligence is[sic]strong enough to be a concern.”);Peter Holley,Apple Co-Founder on Artificial Intelligence:“The Future Is Scary and Very Bad for People,”WASH.POST(Mar.24,2015),http://www.washingtonpost.com/blogs/the-switch/wp/2015/03/24/apple-cofounderon-artificial-intelligence-the-future-is-scaryand-very-bad-for-people/[https://perma.cc/6YRC-QDSG](quoting Steve Wozniak,“If we build these devices to take care of everything for us,eventually they’ll think faster than us and they’ll get rid of the slow humans to run companies more efficiently”).

10 The scholarship on the related field of law and robotics is somewhat betterdeveloped.See generally,e.g.,Ryan Calo,Robotics and the Lessons of Cyberlaw,103 CALIF.L.REV.513(2015);Gabriel Hallevy,“I,Robot-I,Criminal”-When Science Fiction Becomes Reality:Legal Liability of AI Robots Committing Criminal Offenses,22 SYRACUSE SCI.&TECH.L.REP.1(2010);F.Patrick Hubbard,“Sophisticated Robots”:Balancing Liability,Regulation,and Innovation,66 FLA.L.REV.1803(2014).

11 See HENRY M.HART,JR.&ALBERT M.SACKS,THE LEGAL PROCESS:BASICP R O B L E M S I N T H E M A K I N G A N DAPPLICATION OF LAW lx(1994).

12 In this article,the term“autonomous machines”refers to machines that“act independently of direct human instruction,based on information the machine itself acquires and analyzes.”David C.Vladeck,Machines Without Principals:Liability Rules and Artificial Intelligence,89 WASH.L.REV.117,121(2014);see also Matthew U.Scherer,Who’s to Blame(Part 2):What Is an“Autonomous”Weapon?,LAW AND AI(Feb.10,2016),http://www.lawandai.com/2016/02/10/what-is-an-autonomous-weapon/[https://perma.cc/668Q-9VWJ](defining autonomy as the ability of a system to operate free from human direction,monitoring,and control).As with other terms used in this article,see infra notes 46-47,the use of the term“autonomy”is not meant to imply that such machines possess the metaphysical qualities of consciousness or self-awareness.

13 See John Mc Carthy,What is Artificial Intelligence?,JOHN MCCARTHY’S HOMEPAGE 2-3(Nov.12,2007),http://www-formal.stanford.edu/jmc/whatisai.pdf[https://perma.cc/U3RT-Q7JK]

14 Some of these characteristics are,of course,present to various degrees in some other animals as well.Most notably,there is extensive scientific literature examining the cognitive abilities of nonhuman primates and cetaceans.See generally,e.g.,DAVID PREMACK,INTELLIGENCE IN APEAND MAN(1976);Olivier Pascalis&Jocelyne Bachevalier,Face Recognition in Primates:ACross-Species Study,43 BEHAV.PROCESSES87(1998);Rachel Adelson,Marine Mammals Master Math,MONITOR PSYCHOL.Sept.2005,at 22,http://www.apa.org/monitor/sep05/marine.aspx[https://perma.cc/DU3G-4VP8].

15 RUSSELL&NORVIG,supra note 7,at 2

16 A.M.Turing,Computing Machinery and Intelligence,59 MIND 433,442(1950)

17 See NILS J.NILSSON,THE QUEST FORARTIFICIAL INTELLIGENCE 194(2010)(discussing the computer chess program Mac Hack VI’s performance in tournaments against human players in 1967).

18 See BRUCE PANDOLFINI,KASPAROV ANDDEEP BLUE:THE HISTORIC CHESS MATCHBETWEEN MAN AND MACHINE 7-8(1997).

19 See Moshe Y.Vardi,Artificial Intelligence:Past and Future,COMM.ACM,Jan.2012,at 5,5(2012).

20 RUSSELL&NORVIG,supra note 7,at 4.

21 THE AMERICAN HERITAGE DICTIONARYand MERRIAM-WEBSTER’S COLLEGIATEDICTIONARY both direct readers to the entry for“intention”for a list of synonyms of“goal.”Goal,THE AMERICAN HERITAGEDICTIONARY OF THE ENGLISH LANGUAGE(4th ed.2000);Goal,MERRIAM-WEBSTER’SCOLLEGIATE DICTIONARY(11th ed.2003).

22 See infra note 53 and accompanying text.

23 See Neil Johnson et al.,Abrupt Rise of New Machine Ecology Beyond Human Response Time,SCI.REPORTS,Sept.11,2013,at 1,2;Kessler,supra note 1.

24 See,e.g.,Smith&Anderson,supra note 34,at 52.

25 See NATE SILVER,THE SIGNAL AND THENOISE:WHY SO MANY PREDICTIONS FAIL-BUT SOME DON’T 287-88(2012).

26 SILVER,supra note 36,at 287-88.

27 Herbert A.Simon,Rational Choice and the Structure of the Environment,63 PSYCHOL.REV.129,136(1956).

28 See Jonathan Schaeffer et al.,Checkers Is Solved,317 SCI.1518,1518-20(2007).Of course,this only applies to finding solutions to problems that can be formalized and reduced to computer code.

29 See SILVER,supra note 36,at 287-88;Calo,supra note 11,at 532,539(using the term“emergence”to refer to the“unpredictably useful behavior”of robots,and noting that such behavior“can lead to solutions no human would have come to on her own”).

30 See Calo,supra note 11 at 538(“Emergent behavior is a clearly stated goal of robotics and artificial intelligence....”).

31 RESTATEMENT(THIRD)OF TORTS:PHYS.&EMOT.HARM§34 cmt.b(AM.LAW INST.2010).For a general discussion on the issues surrounding liability for harm caused by robots,see WENDELL WALLACH&COLIN ALLEN,MORAL MACHINES:TEACHING ROBOTSRIGHT FROM WRONG 197-214(2009).

32 Here,the term“learning AI”is not meant to imply that the AI system consciously learns,but rather that it is able to gather and,through machine learning,use new data to change how it acts in the world

33 As with“autonomous”and“learning,”the term“experience”is not meant to imply consciousness,but rather to serve as a useful shorthand for the actionable data that an AIsystem gathers regarding its environment and the world in which it exists.

34 See Pei Wang,The Risk and Safety of AI,AGENERAL THEORY OF INTELLIGENCE,https://sites.google.com/site/narswang/EBook/topic-list/the-risk-and-safety-of-ai[https://perma.cc/5LY3-CTLD](“An adaptive system’s behaviors are determined both by its nature(i.e.,initial design)and its nurture(i.e.,postnatal experience).Though it is still possible to give the system certain innate beliefs and motivations,they will not fully determine the system’s behaviors.”).

35 See Jack M.Balkin,The Path of Robotics Law,6 CALIF.L.REV.CIRCUIT 45,52(2015),http://www.californialawreview.org/wp-content/uploads/2015/06/Balkin-Circuit.pdf[https://perma.cc/AP4A-3YX8](“[A]lthough the risk of some kind of injury at some point in the future is foreseeable whenever one introduces a new technology,how and when an injury occurs may not be particularly foreseeable to each of the potential defendants....”).

36 See RESTATEMENT(THIRD)OF TORTS:APPORTIONMENT OF LIABILITY§§10-17(“Liability of Multiple Tortfeasors for Indivisible Harm”);id.§§22-23(“Contribution and Indemnity”);Calo,supra note 11,at 554-55;Balkin,supra note 49,at 53.

37 See Johnson et al.,supra note 33

38 See,e.g.,Luke Muehlhauser&Nick Bostrom,Why We Need Friendly AI,13 THINK 41,41-43(2014);Stuart Russell,Of Myths and Moonshine,EDGE,http://edge.org/conversation/jaron_lanier-the-myth-of-ai#26015[https://perma.cc/PLG8-RWBZ];NATE SOARES&BENJAFALLENSTEIN,MACHINE INTELLIGENCERES.INST.,ALIGNING SUPERINTELLIGENCEWITH HUMAN INTERESTS:A TECHNICALRESEARCH AGENDA 2(2014),https://intelligence.org/files/Technical Agenda.pdf[https://perma.cc/2XQT-NEXV].

39 RUSSELL&NORVIG,supra note 7,at 1037

40 RUSSELL&NORVIG,supra note 7,at 1037;see also RICHARD A.POSNER,CATASTROPHE:RISK AND RESPONSE 41(“Unless carefully programmed,[military]robots might prove indiscriminately destructive and turn on their creators.”).

41 Nick Bostrom coined the term“superintelligence”to refer to the abilities of such a machine.Bostrom defines superintelligence as“an intellect that is much smarter than the best human brains in practically every field,including scientific creativity,general wisdom and social skills.”Nick Bostrom,How Long Before Superintelligence?,NICK BOSTROM’SHOME PAGE,http://www.nickbostrom.com/superintelligence.html[https://perma.cc/7XW2-VLRC].

42 See Johnson et al.,supra note 33,at 1.

43 See,e.g.,Nils Pratley,The Trillion-Dollar Questions over the Flash Crash and the Hound of Hounslow,GUARDIAN(Apr.25,2015,11:00AM),http://www.theguardian.com/business/2015/apr/25/flash-crash-hound-of-hounslow-trilliondollar-question[https://perma.cc/88QE-5FR4]

44 See,e.g.,John O.Mc Ginnis,Accelerating AI,104NW.U.L.REV.1253,1262(2010)

45 See,e.g.,Calo,supra note 11,at 534(“Programming dictates behavior in complex ways.Code interacts with other code and various inputs,for instance,operator instructions or sensor data.”).

46 John Paul Mac Duffie&Takahiro Fujimoto,Why Dinosaurs Will Keep Ruling the Auto Industry,88HARV.BUS.REV.23,23(2010).

47 See Vladeck,supra note 13,at 148(citing the potential for“undetectable failure”in the components of automated driving systems as a drawback to holding manufacturers primarily liable for defects in autonomous vehicles).

48 See,e.g.,Vernor Vinge,The Coming Technological Singularity:How to Survive in the Post-Human Era,10129 NASA CONF.PUBLICATION 11,15(1992),http://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/19940022855.pdf[https://perma.cc/J2SU-UK5E](“In fact,the competitive advantage...of every advance in automation is so compelling that passing laws,or having customs,that forbid[human-level AI]merely assures that someone else will get them first.”).

49 David Vladeck explores these issues by asking who should be held responsible for the damage caused by HAL 9000,the AI villain in Stanley Kubrick’s 2001:A Space Odyssey:Would it be fair to hold liable the companies that designed,programmed,or manufactured HAL 9000,even though they embedded in HAL’s“thinking”systems the first rule of autonomous machinesi.e.,never harm a human-and even though the evidence strongly suggests that HAL“taught”himself to defy their instructions?Or should the creators of machines that have the capacity to“think”be held strictly liable whenever anything goes wrong?If so,on what theory?The theory that the wrongful conduct itself is proof of a defect?Or on an insurance-based theory that the creators are in a better economic position to absorb the cost of the injury than the person harmed?Vladeck,supra note 13,at 125.Vladeck further notes that courts may hesitate to assign liability to automated systems where credible alternative theories of liability exist.See id.At 140 n.78.Vladeck cites Ferguson v.Bombardier Servs.Corp.,where the plaintiffs claimed that a defective autopilot system caused a plane crash.Id.(citing 244 F.App’x 944,947(11th Cir.2007)).The trial court,in a ruling upheld on appeal,excluded a plaintiffs’witness from testifying because the witness’proposed testimony was equally consistent with the defendants’theory that the plane had been improperly loaded by its operators.244 F.App’x at 947-49.

50 See,e.g.,SAIF Corp.v.Allen,881 P.2d 773,782-83(Or.1994)(discussing Oregon’s rules for interpreting“inexact”and“delegative”statutory terms)

51 See RESTATEMENT(THIRD)OF TORTS:APPORTIONMENT OF LIABILITY§§10-17(“Liability of Multiple Tortfeasors for Indivisible Harm”);id.§§22-23(“Contribution and Indemnity”).

52 See,e.g.,IBM Watson,IBM,http://www.ibm.com/smarterplanet/us/en/ibmwatson/[https://perma.cc/5BX9-SWE5];Facebook AI Research(FAIR),FACEBOOK,https://research.facebook.com/ai[https://perma.cc/9UW3-TJ2G];Introducing Project Adam:A New Deep-Learning System,MICROSOFT(July 14,2014),http://research.microsoft.com/apps/video/default.aspx?id=220709&r=1[https://perma.cc/Y2WU-PZV5].

53 June 2015,TOP500,http://www.top500.org/lists/2015/06/[https://perma.cc/Q5P8-R72Q].According to TOP500,the most powerful supercomputer as of November 2014 was China’s Tianhe-2,capable of performing 33.86peta FLOPS,or 3.386 x 1016 FLOPS.Id

54 See Kaushal&Nolan,supra note 5(proposing a“new Manhattan Project”for AI);see also Mc Ginnis,supra note 61,at 1265(proposing a“research project,like those funded by the National Institutes of Health”).

55 Cf.Calo,supra note 11,at 537(noting that in order to resolve the challenges that robotics will present to the legal system“[c]ourts may soften or strengthen existing doctrines,import doctrines across subject matter,or resurrect doctrine long forgotten”).

56 See HART,JR.&SACKS,supra note 12,at lx.

57 This paper focuses on national rather than state or provincial legislatures because of the diffuse and easily transportable nature of AI research.Because of these factors,most regional and local legislatures would be able to regulate only a small fraction of AI research.Consequently,any substantive regulations adopted solely by a single sub-national political unit would not likely have a significant effect on the development and deployment of AI as a whole.Of course,national regulations suffer the same disadvantages when compared to international treaties.But because negotiating and ratifying a comprehensive AItreaty would be exceedingly difficult(witness the failures of the Doha round of trade talks and the Copenhagen climate change conference),the prospects for such a treaty seem remote in the absence of preexisting national regulatory systems or a strong consensus that AI poses a global catastrophic risk.See WALLACH&ALLEN,supra note 45,at 212(pointing out“decisions to regulate research”in some countries may not be supported by the governments of other countries because“[v]alues and social pressures differ from country to country and state to state”).

58 See infra Part III.B.2.

59 See infra Parts III.B.4 and III.C.2.

60 See e.g.,Roscoe Pound,Common Law and Legislation,21 HARV.L.REV.384,406(1908)(“We recognize that legislation is the more truly democratic form of lawmaking.We see in legislation the more direct and accurate expression of the general will.”).

61 See HART,JR.&SACKS,supra note 12,at lxi n.42;Todd Eberly,The Death of the Congressional Committee,BALTIMORE SUN(Nov.27,2011),http://articles.baltimoresun.com/2011-11-

27/news/bs-ed-supercommittee-20111127_1_committee-system-committee-chairscommittee-hearings/2[https://perma.cc/S7C7-65PT].

62 See,e.g.,Eberly,See HART,JR.&SACKS,supra note 12,at lxi n.42;Todd Eberly,The Death of the Congressional Committee,BALTIMORE SUN(Nov.27,2011),http://articles.baltimoresun.com/2011-11-27/news/bs-edsupercommittee-20111127_1_committeesystem-committee-chairs-committee-hearings/2[https://perma.cc/S7C7-65PT].

63 Id.at xciv.

64 See Barton,supra note 91,at 1224.

65 Specialty courts and therapeutic courts represent a limited exception to this rule.To date,however,no such specialty courts that are devoted to cases involving a specific industry or technology have been established.Committee work theoretically should permit legislators to specialize to some degree,but this specialization is limited by the relatively small amount of time that legislators spend on committee work and diluted further by the fact that most legislators are members of multiple committees or subcommittees.

66 See,e.g.,12 U.S.C.§242(2012)(members of the Federal Reserve’s Board of Governors can only be removed“for cause”);42 U.S.C.§7171(2012)(members of Federal Energy Regulatory Commission“may be removed by the President only for inefficiency,neglect of duty,or malfeasance in office”).67James O.Freedman,Expertise and the Administrative Process,28 ADMIN.L.REV.363,373(1976).

68 See RICHARD GLOVER,MURPHY ONEVIDENCE 688-89(14th ed.2015).

69 See,e.g.,WASH.PATTERNJUR YINSTRUCTIONS§1.01(2014)(“The only evidence you are to consider consists of testimony of witnesses and exhibits admitted into evidence....The law is contained in my instructions to you.You must disregard anything the lawyers say that is at odds with the evidence or the law in my instructions.”);MICH.MODEL CRIM.JURYINSTRUCTIONS§2.5(2015)(“Evidence includes only the sworn testimony of witnesses,the exhibits admitted into evidence,and anything else I tell you to consider as evidence.”).

70 The National Institute of Standards and Technology,for instance,publishes voluntary and consensus standards on a variety of topics,including a framework on cybersecurity first drafted in 2014.See Press Release,Nat’l Inst.of Standards and Tech.,NIST Releases Cybersecurity Framework Version 1.0(Feb.12,2014),http://www.nist.gov/itl/csd/launchcybersecurity-framework-021214.cfm[https://perma.cc/6QN3-LYAD].

71 Of course,determining the substantive content of tort standards necessarily involves some degree of policy judgment.Thus,policy considerations are legally relevant when a technology is new and the tort standards governing it are still being developed.But the policy decisions that drive the choice of standards are not among the issues that juries must decide in tort trials.

72 This process could be analogized to the“common enterprise”theory of liability,which Vladeck proposes as a tort liability model for autonomous vehicles and,by extension,other AI systems.See Vladeck,supra note 13,at 149.This proposal avoids the common enterprise phrasing because the problem of discreteness,discussed above in Part II.B.3,means that some of the entities who design the components of an AI system may have no organizational relationship to one another,and thus would not constitute a common enterprise under the usual formulation of that doctrine.See,e.g.,FTC v.E.M.A.Nationwide,Inc.,767 F.3d611,637(6th Cir.2014)(“Courts generally find that a common enterprise exists‘if,for example,businesses(1)maintain officers and employees in common,(2)operate under common control,(3)share offices,(4)commingle funds,and(5)share advertising and marketing.’”)(quoting FTC v.Wash.Data Res.,856 F.Supp.2d 1247,1271(M.D.Fla.2012)).

73 See 21 U.S.C.§355(j)(2012);see also Abbreviated New Drug Application(ANDA):Generics,U.S.FOOD&DRUGADMIN.,http://www.fda.gov/Drugs/Development Approval Process/How Drugsare Developedand Approved/Approval Applications/Abbreviated New Drug Application ANDAGenerics/[https://perma.cc/BC7E-Q28H](lastupdated Apr.14,2016).

74 See RESTATEMENT(THIRD)OF TORTS:APPORTIONMENT OF LIABILITY§§10-17(“Liability of Multiple Tortfeasors for Indivisible Harm”);id.§§22-23(“Contribution and Indemnity”).

75 See How Drugs Are Developed and Approved,U.S.FOOD&DRUG ADMIN.,http://www.fda.gov/Drugs/Development Approval Process/How Drugsare Developedand Approved/default.htm[https://perma.cc/WPM3-D4DA](last updated Aug.18,2015).

76 Cf.WALLACH&ALLEN,supra note 45,at 198(“It may also be in[companies producing and utilizing intelligent machines’]interests to promote a kind of independent legalstatus as agents for these machines(similar to that given corporations)as a means of limiting the financial and legal obligations of those who create and use them.”);Vladeck,supranote 13,at 129(suggesting that one approach to liability for harm caused by autonomous vehicles“would be to hold the vehicle itself responsible,assuming,of course,that the law is willing to confer legal‘personhood’on the vehicle and require the vehicle to obtain adequate insurance”).