Liability for Crimes Involving Artificial Intelligence Systems

262 Pages·2015·2.26 MB·English

by  Gabriel Hallevy (auth.)

Gabriel Hallevy
Liability for
Crimes Involving
Artifi cial
Intelligence
Systems
Liability for Crimes Involving Artificial
Intelligence Systems
ThiSisaFMBlankPage
Gabriel Hallevy
Liability for Crimes
Involving Artificial
Intelligence Systems
GabrielHallevy
FacultyofLaw
OnoAcademicCollege
ISBN978-3-319-10123-1 ISBN978-3-319-10124-8(eBook)
DOI10.1007/978-3-319-10124-8
SpringerChamHeidelbergNewYorkDordrechtLondon
LibraryofCongressControlNumber:2014955453
#SpringerInternationalPublishingSwitzerland2015
Thisworkissubjecttocopyright.AllrightsarereservedbythePublisher,whetherthewholeorpart
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation,broadcasting,reproductiononmicrofilmsorinanyotherphysicalway,andtransmissionor
informationstorageandretrieval,electronicadaptation,computersoftware,orbysimilarordissimilar
methodologynowknownorhereafterdeveloped.Exemptedfromthislegalreservationarebriefexcerpts
inconnectionwithreviewsorscholarlyanalysisormaterialsuppliedspecificallyforthepurposeofbeing
enteredandexecutedonacomputersystem,forexclusiveusebythepurchaserofthework.Duplication
ofthispublicationorpartsthereofispermittedonlyundertheprovisionsoftheCopyrightLawofthe
Publisher’s location, in its current version, and permission for use must always be obtained from
Springer.PermissionsforusemaybeobtainedthroughRightsLinkattheCopyrightClearanceCenter.
ViolationsareliabletoprosecutionundertherespectiveCopyrightLaw.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publicationdoesnotimply,evenintheabsenceofaspecificstatement,thatsuchnamesareexempt
fromtherelevantprotectivelawsandregulationsandthereforefreeforgeneraluse.
While the advice and information in this book are believed to be true and accurate at the date of
publication,neithertheauthorsnortheeditorsnorthepublishercanacceptanylegalresponsibilityfor
anyerrorsoromissionsthatmaybemade.Thepublishermakesnowarranty,expressorimplied,with
respecttothematerialcontainedherein.
Printedonacid-freepaper
SpringerispartofSpringerScience+BusinessMedia(www.springer.com)
Preface
Theideaofliabilityforcrimesinvolvingartificialintelligencesystemshasnotbeen
widely researched yet. Advanced technology makes society face new challenges,
notonlytechnological,butlegalaswell.Theideaofcriminalliabilityinthespecific
context of artificial intelligence systems is one of these challenges that should be
thoroughly explored. The main question is who should be criminally liable for
offenses involving artificial intelligence systems. The answer may include the
programmers,themanufacturers,theusers,and,perhaps,theartificialintelligence
systemitself.
In2010afewarticlesofminewerepublishedintheUSAandAustraliaoncertain
aspects of this issue. These articles explored the specific aspects that seemed to be
important to open up an academic discussion on this issue. The main idea of these
articleswasthatcriminallawisnotsupposedtochangetechnology,butshouldadapt
itselftomoderntechnologicalinsights.Theyalsocalledforthinkingandrethinking
theideaofimpositionofcriminalliabilityuponmachinesandsoftware.Perhaps,no
criminalliabilityshouldbeimposedonmachines,butifbasicdefinitionsofcriminal
lawarenotchanged,thisoddandweirdconsequenceisinevitable.
Dozensofcommentsarrivedforeacharticle,andthetimehascomefornarrow
generalization of this idea. The first generalization of this idea was restricted to
tangiblerobots,whichareequippedwithartificialintelligencesoftwareandcommit
homicideoffensesasspecificoffensesandnotthroughderivativecriminalliability.
Thus,mybookWhenRobotsKillwaspublishedin2013intheUSAbyUPNEand
NortheasternUniversityPress.Althoughthebookisacademic,itmadeanattempt
toaddresswiderpopulationotherthanlegalacademics.
The book was found innovative, and reviews were published in various places
suchastheWashingtonPost,theBostonGlobeandtheChronicleReview.Dozens
ofcommentsarrivedaswell.Someofthesecommentscalledforthefinalandfull
academicgeneralizationofthisissue,notrestrictedtotangiblerobots,notrestricted
tohomicideoffensesandopenedforderivativecriminalliability.Theneedwasfor
anacademicprofessionaltextbooktowardsthisissue,althoughitmaynotaddress
towide population.This bookis the final andfullacademicgeneralization ofthis
issue. The general idea expressed in this book relates to all types of advanced
artificial intelligence systems, including both fully operational and planned
systems,toallmodesofcriminalliability,includingdirectandderivativeliability,
andtoalltypesofoffenses.
v
vi Preface
The reader would find in this book a mature thorough theory towards the
criminal liability for offenses involving artificial intelligence systems based on
the current criminal law in most modern legal systems. The involvement of the
artificialintelligencesystemsintheseoffensesmaybeasperpetrators,accomplices
andmere instrumentsfor thecommissionoftheoffense. Oneofthepointsofthis
book is that, perhaps, no criminal liability should be imposed on technological
systems, at least yet, but if basic definitions of criminal law are not changed, this
oddconsequenceisinevitable.
GabrielHallevy
Contents
1 ArtificialIntelligenceTechnologyandModernTechnological
Delinquency. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 ArtificialIntelligenceTechnology. . . . . . . . . . . . . . . . . . . . . . . . 1
1.1.1 TheRiseofArtificialIntelligenceTechnology. . . . . . . . . 1
1.1.2 OutlinesofArtificialIntelligenceTechnology. . . . . . . . . . 6
1.1.3 DailyUsageofArtificialIntelligenceTechnology. . . . . . . 14
1.2 TheDevelopmentoftheModernTechnologicalDelinquency. . . . 16
1.2.1 TheAversionfromWideUsageofAdvancedTechnology. . . 16
1.2.2 DelinquencybyTechnology. . . . . .. . . . . . . .. . . . . . . .. 21
1.2.3 ModernAnalogiesofLiability. . . . . . . . . . . . . . . . . . . . . 26
2 BasicRequirementsofModernCriminalLiability. . . . . . . . . . . . . . 29
2.1 ModernCriminalLiability. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
2.1.1 TheOffense’sRequirements(InRem). . . . . . . . . . . . . . . 31
2.1.2 TheOffender’sRequirements(InPersonam). . . . . . . . . . 34
2.2 LegalEntities. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
2.2.1 CriminalLiability. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
2.2.2 Punishments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3 ExternalElementInvolvingArtificialIntelligenceSystems.. . . . . .. 47
3.1 TheGeneralStructureoftheExternalElement. . . . . . . . . . . . . . 47
3.1.1 IndependentOffenses. . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.1.2 DerivativeCriminalLiability. . . . . . . . . . . . . . . . . . . . . . 49
3.2 CommissionofExternalElementComponentsbyArtificial
IntelligenceTechnology. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
3.2.1 Conduct. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
3.2.2 Circumstances. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
3.2.3 ResultsandCausation. . . . . . . . . . . . . . . . . . . . . . . . . . . 65
4 PositiveFaultElementInvolvingArtificialIntelligenceSystems. . . . 67
4.1 StructureofPositiveFaultElement. . . . . . . . . . . . . . . . . . . . . . . 67
4.1.1 IndependentOffenses. . . . . . . . . . . . . . . . . . . . . . . . . . . 68
4.1.2 DerivativeCriminalLiability. . . . . . . . . . . . . . . . . . . . . . 70
vii
viii Contents
4.2 GeneralIntentandArtificialIntelligenceSystems. . . . . . . . . . . . 82
4.2.1 StructureofGeneralIntent. . . . . . . . . . . . . . . . . . . . . . . . 82
4.2.2 CognitionandArtificialIntelligenceTechnology. . . . . . . 86
4.2.3 VolitionandArtificialIntelligenceTechnology. . . . . . . . . 93
4.2.4 DirectLiability. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
4.2.5 IndirectLiability. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
4.2.6 CombinedLiabilities. . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
4.3 NegligenceandArtificialIntelligenceSystems. . . . . . . . . . . . . . . 120
4.3.1 StructureofNegligence. . . . . . . . . . . . . . . . . . . . . . . . . . 120
4.3.2 NegligenceandArtificialIntelligenceTechnology. . . . . . 124
4.3.3 DirectLiability. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
4.3.4 IndirectLiability. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
4.3.5 CombinedLiabilities. . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
4.4 StrictLiabilityandArtificialIntelligenceSystems. . . . . . . . . . . . 135
4.4.1 StructureofStrictLiability. . . . . . . . . . . . . . . . . . . . . . . 135
4.4.2 StrictLiabilityandArtificialIntelligenceTechnology. . . . 139
4.4.3 DirectLiability. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
4.4.4 IndirectLiability. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
4.4.5 CombinedLiabilities. . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
5 NegativeFaultElementsandArtificialIntelligenceSystems. . . . . . . 147
5.1 RelevanceandStructureofNegativeFaultElements. . . . . . . . . . 147
5.2 NegativeFaultElementsbyArtificialIntelligenceTechnology. . . 150
5.2.1 InPersonamNegativeFaultElements. . . . . . . . . . . . . . . 150
5.2.2 InRemNegativeFaultElements. . . . . . . . . . . . . . . . . . . 168
6 PunishibilityofArtificialIntelligenceTechnology. . . . . . . . . . . . . . 185
6.1 GeneralPurposesofPunishmentsandSentencing. . . . . . . . . . . . 185
6.1.1 Retribution. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
6.1.2 Deterrence. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
6.1.3 Rehabilitation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
6.1.4 Incapacitation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
6.2 RelevanceofSentencingtoArtificialIntelligenceSystems. . . . . . 210
6.2.1 RelevantPurposestoArtificialIntelligenceTechnology. . . 210
6.2.2 OutlinesforImpositionofSpecificPunishments
onArtificialIntelligenceTechnology. . . . . . . . . . . . . . . . 212
Conclusion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
Cases. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
Bibliography. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
Index. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255
1
Artificial Intelligence Technology
and Modern Technological Delinquency
Contents
1.1 ArtificialIntelligenceTechnology........................................................ 1
1.1.1 TheRiseofArtificialIntelligenceTechnology.................................. 1
1.1.2 OutlinesofArtificialIntelligenceTechnology................................... 6
1.1.3 DailyUsageofArtificialIntelligenceTechnology.............................. 14
1.2 TheDevelopmentoftheModernTechnologicalDelinquency.......................... 16
1.2.1 TheAversionfromWideUsageofAdvancedTechnology..................... 16
1.2.2 DelinquencybyTechnology...................................................... 21
1.2.3 ModernAnalogiesofLiability................................................... 26
1.1 Artificial Intelligence Technology
1.1.1 The Rise of Artificial Intelligence Technology
Artificialintelligencetechnologyisthebasisforgrowingnumberofsciencefiction
compositions, such as books and movies. Some of them reflect fears from this
technologyandsomereflecttheenthusiasmtowardsit.Themajorepistemological
question has always remained whether machines can think. Some agree that they
can“think”,butthequestioniswhethertheycanthink(withoutcommas).
The modern answer to this question may be proposed by artificial intelligence
technology.1 This technology is considered to be modern, but its roots are not
necessarilymodern.Infact,sincetheverydawnofhumanitymankindhasalways
soughttoolstoeasedailylife.IntheStoneAge,thesetoolsweremadeofstone.As
mankind discovered the advantages of metal, these tools were made of metal. As
human knowledge became wider, more and more tools were invented to take
growingrolesinhumandailylife.
1For the technical review of this issue and the historical developments see GABRIEL HALLEVY,
WHENROBOTSKILL–ARTIFICIALINTELLIGENCEUNDERCRIMINALLAW1–37(2013).
#SpringerInternationalPublishingSwitzerland2015 1
G.Hallevy,LiabilityforCrimesInvolvingArtificialIntelligenceSystems,
DOI10.1007/978-3-319-10124-8_1