A Government of Laws and Not of Machines

Berman, Emily

doi:10.2139/ssrn.3098995

Cited by 6 publications

(6 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Algorithmic systems can only be as good as the corpus of data on which they are based and, as such, data that is biased along with race, gender, sex, and socioeconomic lines will lead to biased results (Bridges, 2017;Caliskan et al, 2017;Katyal, 2019;Noble, 2018;O'Neil, 2016;Sühr et al, 2021). Although it is true that some of these concerns may be ameliorated with design changes, algorithms' deficiencies may threaten the legitimacy of authorities in democratic societies (Berman, 2018).…”

Section: The Functions and Risks Of Algorithmic Decision-makingmentioning

confidence: 99%

“…They are “black boxes” with little accountability (Innerarity, 2021; Pasquale, 2014). Citron (2007), Crawford and Schultz (2014), and Berman (2018) argue that algorithmic decision-making threatens due process. And, as de Laat (2019) has argued, algorithmic systems that are used to predict human behavior are polypanoptic in the Foucauldian sense: they subject us to surveillance, encourage a race to the norm, and undermine independence.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Governing algorithmic decisions: The role of decision importance and governance on perceived legitimacy of algorithmic decisions

Waldman

Martin

2022

Big Data & Society

View full text Add to dashboard Cite

The algorithmic accountability literature to date has primarily focused on procedural tools to govern automated decision-making systems. That prescriptive literature elides a fundamentally empirical question: whether and under what circumstances, if any, is the use of algorithmic systems to make public policy decisions perceived as legitimate? The present study begins to answer this question. Using factorial vignette survey methodology, we explore the relative importance of the type of decision, the procedural governance, the input data used, and outcome errors on perceptions of the legitimacy of algorithmic public policy decisions as compared to similar human decisions. Among other findings, we find that the type of decision—low importance versus high importance—impacts the perceived legitimacy of automated decisions. We find that human governance of algorithmic systems (aka human-in-the-loop) increases perceptions of the legitimacy of algorithmic decision-making systems, even when those decisions are likely to result in significant errors. Notably, we also find the penalty to perceived legitimacy is greater when human decision-makers make mistakes than when algorithmic systems make the same errors. The positive impact on perceived legitimacy from governance—such as human-in-the-loop—is greatest for highly pivotal decisions such as parole, policing, and healthcare. After discussing the study’s limitations, we outline avenues for future research.

show abstract

Section: The Functions and Risks Of Algorithmic Decision-makingmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Governing algorithmic decisions: The role of decision importance and governance on perceived legitimacy of algorithmic decisions

Waldman

Martin

2022

Big Data & Society

View full text Add to dashboard Cite

show abstract

“…This has led to specific regulatory suggestions to force Facebook’s programmers to modify their code (Grafanaki 2017 , 58). As this simple yet important example shows, the proposed regulatory approach can be enforced through human laws because the law can impose obligations on companies, programmers and users (Berman 2018 , 1330; Sloan and Warner 2019 ). Regulations can include more specific steps such as mandating the use of reinforcement learning to prevent AI machines “from learning the wrong thing… either by a person or the environment” (Arnold and Scheutz 2018 , 62).…”

Section: A Comprehensive Regulatory Approachmentioning

confidence: 99%

Towards an effective transnational regulation of AI

Gervais

2021

AI & Soc

View full text Add to dashboard Cite

Law and the legal system through which law is effected are very powerful, yet the power of the law has always been limited by the laws of nature, upon which the law has now direct grip. Human law now faces an unprecedented challenge, the emergence of a second limit on its grip, a new “species” of intelligent agents (AI machines) that can perform cognitive tasks that until recently only humans could. What happens, as a matter of law, when another species interacts with us, can be integrated into human minds and bodies, makes “real-world” decisions—not through human proxies, but directly—and does all this “intelligently”, with what one could call autonomous agency or even a “mind” of its own? The article starts from the clear premise that control cannot be exercised directly on AI machines through human law. That control can only be effected through laws that apply to humans. This has several regulatory implications. The article’s first discusses what, in any attempt to regulate AI machines, the law can achieve. Having identified what the law can do, the article then canvases what the law should aim to achieve overall. The article encapsulate its analysis in a list of both doctrinal and normative principles that should underpin any regulation aimed at AI machines. Finally, the article compares three transnational options to implement the proposed regulatory approach.

show abstract

“…Numerous studies in this cluster also focus on the relationship between artificial intelligence and law. Berman (2018), for example, examined whether reliance on artificial intelligence for policymaking and implementation is consistent with rule of law. Relatedly, Levendowski (2018) argued for a relook at copyright law to make otherwise excluded material available to computer algorithms and thereby reduce bias in artificial intelligence.…”

Section: Topics Analyzed Using Big Data In the Policy Sciencesmentioning

confidence: 99%