ImaginethatEric,ayoungman,hasbeenconvictedofagang-relatedcrime:hewasfoundbypoliceatthesceneofarobberycarriedoutbyhisfriends.Thesentencingjudgenowneedstomakeadecision.NotknowingwhetherEricposesarisktothepublic,sheturnstoanalgorithmicriskassessment-asetofrules,developedonthebasisofcorrelationsbetweenindividualcharacteristicsandcriminalactivity,whichpredictsthelikelihoodofrecidivism.Ericisaconscientiouscitizen,whohasneverbeforebeenintroublewiththelaw.However,hewasraisedinfostercare,inanareawithhighratesofcrime,poverty,andresidentialinstability-factstowhichthealgorithmattributesahighriskscore.Thejudgerecommendsasentenceofthemaximumpossibleduration,withextendedpost-releasesupervision.Whydoesthismatter?Theanswermostoftengivenisthatitmattersbecauseofinequality,capturedthroughthelanguageof’bias’or’discrimination’:theeffectofusingalgorithmscanbetoexacerbateunjustifieddifferencesbetweenpeople,onthebasisofconsiderationssuchasrace,sex,orsocio-economiccircumstance.Usingadiversesetofcasestudies,andmakingclearpolicyrecommendationsalongtheway,ArtificialJusticeunpacksthereasonsthatwemighthavetoobjecttotheuseofstatisticalalgorithmstoallocatetheburdensofpolicydecisions.Itarguesthatthesereasonsextendbeyondegalitarianconcerns.Importantly,theyincludereasonsthatstemfromthevalueofindividualchoice-ofhavingthechancetoaffectwhathappenstousbychoosingappropriatelyandbeingequippedtoexercisethosechoiceswell.Thebookexploresthesubstantivereasonsthatcontributetoapictureofwhat’satstakeforindividualswhenweusestatisticalalgorithmstomakedecisionsabouthowtotreatothers,andmakesrobustpolicyrecommendationsaboutthescopeandnatureofhuman-algorithminteraction.ArtificialJusticeisacompellingandaccessibletext,whichoffersagreatdealtoawideandinterdisciplinaryaudienceofacademics,students,andthoseotherwiseinterestedinlearningaboutalgorithmicjustice.
