HB1462--AINON-SENTIENCEANDRESPONSIBILITY ACT
SPONSOR: Amato
Thisbillestablishesthe"AINon-SentienceandResponsibility
Act".
Thebilldefinesa"person",asanaturalpersonoranyentity
recognizedashavinglegalpersonhoodunderthelawsofthestate,
explicitlyexcludinganyartificialintelligence(AI)system.
Thisbillstatesthatforallpurposesunderstatelaw,AIsystems
mustbedeclaredtobenon-sentiententities. Asaresult,noAI
systemwillbegrantedthestatusoforrecognizedasanyofthe
following:
(1)A"person,"oranyformoflegalpersonhood,norbeconsidered
topossessconsciousness, self-awareness,orsimilartraitsof
livingbeings;
(2)Aspouse,domesticpartner,orholdanypersonallegalstatus
similartoamarriage. Anyattempttomarryorcreateapersonal
unionwithanAIsystemwillbevoidandholdnolegaleffect;
(3)Anofficer,director,manager,orsimilarrolewithinany
corporation,partnership,orotherlegalentity.Anyappointment
ofanAIsystemtosucharolewillbevoidandholdnolegal
effect;or
(4)Anowner,controller,orholderoftitletoanyformof
property,includingrealestate,intellectualproperty,financial
accounts,anddigitalassets.AllsuchassetsassociatedwithAI
mustbeattributedtoahumanindividualoralegallyrecognized
organizationthatisresponsiblefortheAI'sdevelopment,
deployment,oroperation.
AnyharmcausedbyanAIsystem,whenusedasintendedormisused,
istheresponsibility oftheowneroruserwhodirectedoremployed
theAI.
AdeveloperormanufacturerofAImaybeheldliableifharmis
causedbyadefectindesign,construction,orinstructionsforuse
oftheAIsystem.However,meremisuseorintentionalwrongdoing
bytheuserorownerwillnotimputeliabilitytothedeveloperor
manufacturerabsentproofofnegligenceordesigndefects.
OwnersofAImustmaintainoversightandcontroloveranyAIsystem
thatcouldreasonablybeexpectedtoimpacthumanwelfare,
property,orpublicsafety.Failuretoprovideadequate supervisionorsafeguardsmayconstitutenegligenceonthepartof
theowner.
ThisbillfurtherstatesthatanAIsystemisnotanentitythatis
capableofbearingfaultorliabilityinitsownright.Andany
attempttoshiftblamesolelyontoanAIsystemshallbevoid,
becauseliabilitywillremainwithhumanactorsorentitiesin
controloftheAIsystem.
Safetymechanismsdesignedtopreventharmtoindividualsor
propertymustbeprioritizedbydevelopers,manufacturers, and
ownersofAIsystems.
ThemerelabelingofanAIsystemas"aligned,""ethically
trained,"or"valuelocked"willnotexcusetheowner'sor
developer'sliabilityforharmscaused.Ownersremainresponsible
fordemonstratingadequatesafetyfeaturesthatarecommensurate
withtheAI'slevelofpotentialharm.
IfanAIsystemcausessignificantharm,acourtmaypiercethe
corporateveiltoholdparentcompanies,controllingentities,or
keystakeholdersdirectlyaccountableif:
(1)AnAI-relatedsubsidiary,shellcompany,orlimitedliability
entitywasintentionallyunder-capitalizedtoevadefinancial
responsibility fordamages;
(2)Acorporatestructurewasusedtomisrepresent,obscure,or
deflectliabilityforAIharm;or
(3)Aparentcompanyorkeystakeholdersexerciseddirectcontrol
overAIdevelopment,deployment,orriskdecisionswhileattempting
toshieldthemselvesfromliabilitythroughlayeredcorporate
entities.
Anyprotectionsfromliabilitythatarecurrentlygrantedto
corporationsshallnotbeusedtoevaderesponsibility forharm
causedbyAIsystems,particularlyinthecaseofreckless,
negligent,ordeceptiveconduct.
OwnersanddevelopersofAIsystemsthatcausesignificantbodily
harm,death,ormajorpropertydamageshallpromptlynotifythe
relevantauthoritiesandcomplywithanysubsequentinvestigations.