ARTIFICIALINTELLIGENCEANDLIFEIN2030ONEHUNDREDYEARSTUDYONARTIFICIALINTELLIGENCE|REPORTOFTHE2015STUDYPANEL|SEPTEMBER2016PREFACETheOneHundredYearStudyonArtificialIntelligence,launchedinthefallof2014,isalong-terminvestigationofthefieldofArtificialIntelligence(AI)anditsinfluencesonpeople,theircommunities,andsociety.Itconsidersthescience,engineering,anddeploymentofAI-enabledcomputingsystems.Asitscoreactivity,theStandingCommitteethatoverseestheOneHundredYearStudyformsaStudyPaneleveryfiveyearstoassessthecurrentstateofAI.TheStudyPanelreviewsAI’sprogressintheyearsfollowingtheimmediatelypriorreport,envisionsthepotentialadvancesthatlieahead,anddescribesthetechnicalandsocietalchallengesandopportunitiestheseadvancesraise,includinginsucharenasasethics,economics,andthedesignofsystemscompatiblewithhumancognition.TheoverarchingpurposeoftheOneHundredYearStudy’speriodicexpertreviewistoprovideacollectedandconnectedsetofreflectionsaboutAIanditsinfluencesasthefieldadvances.Thestudiesareexpectedtodevelopsynthesesandassessmentsthatprovideexpert-informedguidancefordirectionsinAIresearch,development,andsystemsdesign,aswellasprogramsandpoliciestohelpensurethatthesesystemsbroadlybenefitindividualsandsociety.1TheOneHundredYearStudyismodeledonanearliereffortinformallyknownasthe“AAAIAsilomarStudy.”During2008-2009,thethenpresidentoftheAssociationfortheAdvancementofArtificialIntelligence(AAAI),EricHorvitz,assembledagroupofAIexpertsfrommultipleinstitutionsandareasofthefield,alongwithscholarsofcognitivescience,philosophy,andlaw.Workingindistributedsubgroups,theparticipantsaddressednear-termAIdevelopments,long-termpossibilities,andlegalandethicalconcerns,andthencametogetherinathree-daymeetingatAsilomartoshareanddiscusstheirfindings.Ashortwrittenreportontheintensivemeetingdiscussions,amplifiedbytheparticipants’subsequentdiscussionswithothercolleagues,generatedwidespreadinterestanddebateinthefieldandbeyond.TheimpactoftheAsilomarmeeting,andimportantadvancesinAIthatincludedAIalgorithmsandtechnologi...
