CONTENTS&ABSTRACTS
jectivity,relevance,andlegalitymustbeexamined.Largelanguagemodelevidencereflectsobjectivefactsin
itscontent,isperceptiblyobjectiveinitsform,demonstratingobjectivity.Itisdeeplyintegratedintosocial
life,connectingwiththefactsofcasesinvariousscenarios,suchascivil,criminal,andadministrativecases,
exhibitingrelevance.Whileitcurrentlylacksalegalfoundation,itcanacquirelegalitythroughtheenhance-
mentofthelegalframework.Therefore,thequestionofwhetherlargelanguagemodelevidencequalifiesas
evidencerequiresspecificconsiderationinfuturecases.Forthesakeoflitigationefficiency,thepresentationof
largelanguagemodelevidencemayvaryindifferentlitigationscenarios.Instandardcases,onlyhumanꢄma-
chineinteractionmaterialsandtheuser’slocalcontextinformationmaysuffice,whilesignificantcasesmay
requireacomprehensivepresentationoflargelanguagemodelevidence.
Withtheadvancementofcuttingꢄedgetechnology,newtypesoftechnologicalevidencehaveemerged,in-
cludingbigdataevidenceandartificialintelligenceevidence.Bigdataevidenceandlargelanguagemodelevi-
dencefacesimilarchallengesrelatedtothe“blackbox”,admissibility,andcategorization.Additionally,big
dataevidencehasbeenwidelyutilizedinpractice,makingtheassociatedtheoriesavaluablereferenceforthe
developmentoflargelanguagemodelevidence.Whileartificialintelligenceevidenceisahigherꢄlevelconcept
thanlargelanguagemodelevidence,theextensivediversityinthefieldofartificialintelligencetechnologyim-
pliesthatthefindingsofAIevidenceresearchmaynotnecessarilybedirectlyapplicabletolargelanguage
modelevidence.
Largelanguagemodelevidenceexhibitsthefollowingcharacteristics.Firstly,itpossessesahighdegree
ofintuitiveness,asmaterialspertainingtohumanꢄmachineinteractioninlargelanguagemodelevidencecanbe
readilyperceivedbyhumans.Secondly,ithaslimitedinterpretabilityduetoinherentblackꢄboxeffectsande-
mergentcharacteristics,resultinginitsrelativelyweakinterpretability.Thirdly,theremaybeevidencebiasin
largelanguagemodelevidence,potentiallyfavoringaselectfewtechnologycompanies.Ourlegalsystemcan
drawinspirationfromtheelectronicdatadisclosuresystemsintheAngloꢄAmericanlegaltraditiontoestablish
disclosureobligationsfortechnologycompanies.Finally,ithaslimitedidentifiability,ashumansmaynotnec-
essarilybeabletodiscernwhetherapieceofmaterialwasgeneratedbyalargelanguagemodelintheabsence
ofspecificindicators.Thisissuecanpotentiallybemitigatedthroughthedeepsyntheticregulation.
TheTripleLogicoftheSystematizationofBlockchainEvidenceRulesandIts
System Development
DuanLupingꢃ74ꢀ83
AlthoughthePeople’sCourtOnlineLitigationRulespreliminarilyregulatedthereviewandidentifica-
tionrulesofblockchainevidencewithfourspecialarticles,howtoestablishasystematicframeworkthatena-
blesjudgestosubjectivelydaretomakedeterminationsobjectivelyaccuratelyidentifyblockchainevidence,
,
andatthesametimeeffectivelysafeguardthelitigants’righttolitigation,isstillworthcontemplating.From
theperspectiveoftheactualmotivationandtheoreticaldemandsoftheintegrationofscienceandtechnology
intothejudiciary,theblockchainevidencerulesinthedigitalagemustconsideratleastthefollowingthree
levelsofprogressivelogic,namelytherealguaranteelogicbasedontheblockchaintrustmechanism,thelogic
ofdifferentialtreatmentbasedonscientificcognitionofthetechnicalcharacteristicsoftheblockchain,andthe
logicofrightsprotectionbasedontherequirementsofdigitaldueprocess.Basedonthistriplelogic,itispos-
sibletosystematicallyconstructrulesforthepresumptionoftheauthenticityofblockchainevidence,accu-
ratelygraspthetechnicalprinciples,setblockchainevidencereviewandidentificationstandards,andprotect
203