Seite - 537 - in Menschenrechte und Gerechtigkeit als bleibende Aufgaben - Beiträge aus Religion, Theologie, Ethik, Recht und Wirtschaft

© 2020, Vandenhoeck & Ruprecht GmbH & Co. KG, Göttingen ISBN Print: 9783847111658 – ISBN E-Lib: 9783737011655 eryone ineveryhousehold”.’9 In thehandsof authoritariangovernments these powerful technologies amplify existing threats to freedomand equality. How- ever, the ethical threats these technologiespose transcendtheirpotentially im- moralusebyauthoritariangovernments.Rather,theextentofdatacollected,the facility tomodel and triangulate, together with their predictive capacity also posesignificant risks tocivil andpolitical liberties indemocratic societies, and inparticular tomarginalised individualsandgroups.Asnotedby theBerkman KleinCentre in its assessmentof ‘Artificial Intelligence andHumanRights’ ‘in relationtotherightstoequality,freeexpression,association,assemblyandwork (…)the impactofAI (…)hasbeenmorenegative thanpositive todate.’10 The predictive capacity of deep learning algorithms raises further ethical challenges.Predictivetechnologiesarealreadyextensivelydeployedinfinancial, health and criminal justice systems andassessments of their equity relative to traditionalhuman-basedsystemsvary.Forexample,TheBerkmanKleinCentre report notes anumberof arenaswhere thedeployment ofAIhashadpositive outcomes.Regardingaccess to the financial systemithas foundthat ‘compared to the status quo credit scoring algorithms, the introduction of AI into the lending process is likely to have an overall positive impact on the ability of low-incomeborrowers toaccesscredit’.This isbecause thewidevarietyofdata sources used seems to ‘improve the ability of well-qualified individuals from marginalisedcommunities toaccesscredit.’11However,determining theethical and human rights impacts of the predictive dimension of AI technologies is complex, inpartbecause these technologies arebeing implemented inexisting socialinstitutionsthatalreadydiscriminatetoagreaterorlesserextent.Afurther complexity relates to the nature, quality and extent of the data onwhich the predictive capacity of AI depends. InWeapons ofMathDestruction: HowBig Data Increases InequalityandThreatensDemocracy, CathyO’Neill argues that predictive analytics based on algorithms tend to punish the poor.12 She uses algorithmichiringpracticesasoneexample,butheranalysisextendstoarange of social andeconomiccontexts, arguing that inmost cases, onaccountof the partial, discriminatory anddecontextualizednatureof thedataonwhichAI is trained,ortrainsitself,it isinevitablethatmachinelearningaccentuatesexisting biasesandexclusions.Moreover, shepointsout thatasAI is embedded further 9 ZakDoffman,ChinaIsUsingFacialRecognitionToTrackEthnicMinorities,EvenInBeijing, in:ForbesMay3/2019. 10 FilippoRaso /HannahHilligoss /VivekKrishnamurthy /ChristopherBavitz / LevinKim, Artificial Intelligence andHumanRights, Opportunities and Risks,CambridgeMA., Be- rkmanKleinCentre2018. 11 Ibid., p. 31. 12 Cathy O’ Neill, Weapons of Math Destruction: How Big Data Increases Inequality and ThreatensDemocracy,USA2016. DesigninganEthicalFuture:CanArtificial IntelligenceHelp? 537 Open-Access-Publikation im Sinne der CC-Lizenz BY 4.0