Accountability of Algorithms in the GDPR and beyond: A European Legal Framework on Automated Decision-Making

Castets-Renard, Céline

doi:10.2139/ssrn.3391266

Cited by 10 publications

(10 citation statements)

References 1 publication

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Further validation of the risk profiling system will be provided by sharing with the GPs the information collected through the Dress-KINESIS (baseline data, risk level, and follow-up data) about the GPs’ patients who provided explicit consent. This validation limits the criticisms correlated to a mere algorithmic decision-making process as bias, opacity, and risk of discrimination [ 39 ], according to the Articles 4 and 22 and the Recital 71 of the GDPR about profiling.…”

Section: Resultsmentioning

confidence: 99%

Facing the National Recovery and Resilience Plan: Sources of Data, Indicators, and Participatory Strategies in Healthcare and Social Fields

Franchini

Molinaro

Caiolfa³

et al. 2021

IJERPH

View full text Add to dashboard Cite

Innovation in governance and services should be the target of the Italian National Recovery and Resilience Plan. Monitoring processes, impacts, and outcomes requires a system of new indicators that are practical to collect. Secondary data sources, their availability, and their information potential should be evaluated, and primary sources should be implemented to supplement traditional disease surveillance. This work highlights the most relevant aspects for bridging the mismatching between complex community needs and current health/social supply and how those aspects could be faced. As a result, we propose a structured multi-phases process for setting the design and functionalities of a cooperative information system, built on the integration between secondary and primary data for informing policies about chronic low back pain (CLBP), a widely recognized determinant of disability and significant economic burden. In particular, we propose the Dress-KINESIS, a tool for improving community capacity development and participation that allows one to freely collect big health and social data and link it to existing secondary data. The system also may be able to monitor how the resources are distributed across different care sectors and suggest how to improve efficiency based on the patient’s CLBP risk stratification. Moreover, it is potentially customizable in other fields of health.

show abstract

Section: Resultsmentioning

confidence: 99%

Facing the National Recovery and Resilience Plan: Sources of Data, Indicators, and Participatory Strategies in Healthcare and Social Fields

Franchini

Molinaro

Caiolfa³

et al. 2021

IJERPH

View full text Add to dashboard Cite

show abstract

“…The increasing availability of data and sophistication of algorithms (including the rebirth of machine learning / neural networks [19]) has enabled more uses and misuses of algorithmically controlled, automated decision-making (ADM, for short) [15]. The scaling of such innovations has happened in a context where algorithms, which still bear little accountability [8], [20], are involved in automating news recommendations [21], [22], advertising [23], but also work [24], [25][26] and other highly sensitive processes such as social ranking, crime prediction, and bail, parole and criminal sentencing [27]- [29] .…”

Section: Automated Decision-making (Adm)mentioning

confidence: 99%

“…Considering this new reality, some efforts have been made to regulate ADM, such as article 22 of the General Data Protection Regulation (GDPR) and Article 29 Working Party's Guidelines on automated decisionmaking and profiling in Europe. However, these existing frameworks remain far from perfect at adequately addressing problems of opacity and discrimination related to machine learning processing and the explanations of automated decision-making [8].…”

Section: Introductionmentioning

confidence: 99%

Algorithmically Controlled Automated Decision-Making and Societal Acceptability: Does Algorithm Type Matter?

Grangé¹

2022

Proceedings of the Annual Hawaii International Conference on System Sciences

View full text Add to dashboard Cite

As technological capabilities expand, an increasing number of decision-making processes (e.g., rankings, selections, exclusions) are being delegated to computerized systems. In this paper, we examine the societal acceptability of a consequential decisionmaking system (university admission) to those subject to the decision (i.e., applicants). We analyze two key drivers: the nature of the decision-making agent (a human vs an algorithm) and the decision-making logic used by the agents (predetermined vs emerging). Consistent with uniqueness neglect theory, we propose that applicants will be more positive toward the use of human agents compared to computerized systems. Consistent with the theory of procedural justice, we further argue that applicants will find the use of a predetermined logic to be more acceptable than an emerging logic. We present the details and results of a factorial survey designed to test our theoretical model.

show abstract

“…Critics claim that the legal framework provides too much freedom to data controllers and insufficiently protects individuals [94]. Furthermore, as ADM systems are very complex, the information should be presented in a comprehensible manner for each individual, and the system's "intentions" made clear [94]. What is problematic is that, not all parties involved have a right to an explanation, for instance, the general public.…”

Section: Human-centric Adm Systems For Credit Scoringmentioning

confidence: 99%

“…What is problematic is that, not all parties involved have a right to an explanation, for instance, the general public. Providing the public with information concerning the ADM systems' functioning, however, would be beneficial to reduce public concerns and to improve individuals' understanding of the use of their data [94]. By increasing knowledge about these systems, social accountability could be created, and influence could be exercised to make these systems more human-centric.…”

Section: Human-centric Adm Systems For Credit Scoringmentioning

confidence: 99%

Trust, Automation Bias and Aversion: Algorithmic Decision-Making in the Context of Credit Scoring

Gsenger¹,

Strle²

2021

Interdisciplinary Description of Complex Systems

View full text Add to dashboard Cite

Algorithmic decision-making (ADM) systems increasingly take on crucial roles in our technology-driven society, making decisions, for instance, concerning employment, education, finances, and public services. This article aims to identify peoples' attitudes towards ADM systems and ensuing behaviours when dealing with ADM systems as identified in the literature and in relation to credit scoring. After briefly discussing main characteristics and types of ADM systems, we first consider trust, automation bias, automation complacency and algorithmic aversion as attitudes towards ADM systems. These factors result in various behaviours by users, operators, and managers. Second, we consider how these factors could affect attitudes towards and use of ADM systems within the context of credit scoring. Third, we describe some possible strategies to reduce aversion, bias, and complacency, and consider several ways in which trust could be increased in the context of credit scoring. Importantly, although many advantages in applying ADM systems to complex choice problems can be identified, using ADM systems should be approached with caree.g., the models ADM systems are based on are sometimes flawed, the data they gather to support or make decisions are easily biased, and the motives for their use unreflected upon or unethical.

show abstract

Accountability of Algorithms in the GDPR and beyond: A European Legal Framework on Automated Decision-Making

Cited by 10 publications

References 1 publication

Facing the National Recovery and Resilience Plan: Sources of Data, Indicators, and Participatory Strategies in Healthcare and Social Fields

Facing the National Recovery and Resilience Plan: Sources of Data, Indicators, and Participatory Strategies in Healthcare and Social Fields

Algorithmically Controlled Automated Decision-Making and Societal Acceptability: Does Algorithm Type Matter?

Trust, Automation Bias and Aversion: Algorithmic Decision-Making in the Context of Credit Scoring

Contact Info

Product

Resources

About