scholarly journals Preventing Racial Bias in Federal AI

2020 ◽  
Vol 16 (02) ◽  
Author(s):  
Morgan Livingston
Keyword(s):  
Artificial Intelligence ◽  
Racial Bias ◽  
Racial Diversity ◽  
Federal Agencies ◽  
Racial Groups ◽  
Regulatory Enforcement ◽  
Racial Inequalities ◽  
Administrative Functions ◽  
Oecd Principles ◽  
Us Federal Government

Artificial Intelligence (AI) systems are increasingly used by the US federal government to replace or support decision making. AI is a computer-based system trained to recognize patterns in data and to apply these patterns to form predictions about new data for a specific task. AI is often viewed as a neutral technological tool, bringing efficiency, objectivity and accuracy to administrative functions, citizen access to services, and regulatory enforcement. However, AI can also encode and amplify the biases of society. Choices on design, implementation, and use can embed existing racial inequalities into AI, leading to a racially biased AI system producing inaccurate predictions or to harmful consequences for racial groups. Racially discriminatory AI systems have already affected public systems such as criminal justice, healthcare, financial systems and housing. This memo addresses the primary causes for the development, deployment and use of racially biased AI systems and suggests three responses to ensure that federal agencies realize the benefits of AI and protect against racially disparate impact. There are three actions that federal agencies must take to prevent racial bias: 1) increase racial diversity in AI designers, 2) implement AI impact assessment, 3) establish procedures for staff to contest automated decisions. Each proposal addresses a different stage in the lifecycle of AI used by federal agencies and helps align US policy with the Organization for Economic Co-operation and Development (OECD) Principles on Artificial Intelligence.

scholarly journals Developmental Origins of Anti-Black Bias in White US Children: Exposure to and Beliefs about Racial Inequality

2021 ◽  
Author(s):  
Michael Rizzo ◽  
Tobias Britton ◽  
Marjorie Rhodes
Keyword(s):  
United States ◽  
Early Childhood ◽  
Racial Inequality ◽  
Racial Bias ◽  
Social Systems ◽  
The United States ◽  
Racial Groups ◽  
Developmental Origins ◽  
Structural Racism ◽  
Racial Inequalities

Anti-Black racism remains a pervasive crisis in the United States today. Racist social systems are rooted in prejudicial beliefs that reinforce and perpetuate racial inequalities. These beliefs have their developmental origins in early childhood and are difficult to change once entrenched in adolescence and adulthood. What causes children to form prejudicial beliefs and racial biases—and what steps can be taken to preempt them from forming—remain open questions. Here we show that children’s exposure to and beliefs about racial inequalities predict the formation of anti-Black biases in a sample of 712 White children (4-8 years) living across the United States. Drawing from constructivist theories in developmental science, we outline a novel account of the emergence of racial bias in early childhood: As children observe racial inequalities in the world around them, they develop beliefs about the causal factors underlying those inequalities. Children who believe that inequalities reflect the inherent superiority/inferiority of racial groups develop biases that perpetuate this worldview, whereas those who recognize the extrinsic causes of racial inequalities develop attitudes geared towards rectification. Our results demonstrate the importance of early intervention to disrupt problematic beliefs before they emerge and highlight children’s awareness of structural racism as an important target for anti-racist intervention.

scholarly journals Beliefs About Social Norms and Racial Inequalities Predict Variation in the Early Development of Racial Bias

2020 ◽  
Author(s):  
Michael Rizzo ◽  
Emily Green ◽  
Yarrow Dunham ◽  
Emile Bruneau ◽  
Marjorie Rhodes
Keyword(s):  
Native American ◽  
Social Norms ◽  
Racial Bias ◽  
Middle Eastern ◽  
Developmental Trajectory ◽  
Racial Groups ◽  
Racial Inequalities ◽  
Racial Biases ◽  
Interracial Friendships ◽  
Over Time

Racial biases emerge early in development and threaten the wellbeing of members of marginalized racial groups. We examined how two psychological mechanisms—normative beliefs about interracial friendships and explanatory beliefs about racial inequalities—developmentally relate to the emergence of racial bias over time in a longitudinal study with a sample of 4-year-old children (N=116; 59 female, 57 male; 56 White, 19 Asian, 8 Hispanic/Latinx, 6 Black/African-American, 6 bi- or multiracial, 2 Middle-Eastern, 1 Native American, 18 not-specified; living in upper-middle class, urban, neighborhoods). Children were interviewed 3 times over a 6-month period during the school year. During each wave, children completed assessments of their Black/White racial biases (Interracial Attitudes, Playmate Preferences, Ambiguous Attributions), beliefs about social norms (Parent Norms, Peer Norms), and explanations for racial inequalities. Over the 6-month study, beliefs that parents and peers did not value interracial friendships related to increased racial biases in and across time. Further, endorsement of essentialist beliefs about racial inequalities predicted the developmental trajectory of racial bias; children who initially endorsed essentialized beliefs about racial inequalities developed higher levels of racial biases over time. Results speak to the importance of early and persistent intervention efforts targeting foundational beliefs about the social world to preempt the development of racial bias.

Racial Order, Racialized Responses: Interminority Politics in a Diverse Nation

2021 ◽  
Author(s):  
Efrén O. Pérez ◽  
E. Enya Kuo
Keyword(s):  
Survey Data ◽  
Racial Diversity ◽  
Two Dimensions ◽  
People Of Color ◽  
Racial Groups ◽  
Demographic Changes ◽  
Unified Framework ◽  
Political Responses ◽  
Original Survey ◽  
Parallel Growth

America's racial sands are quickly shifting, with parallel growth in theories to explain how varied groups respond, politically, to demographic changes. This Element develops a unified framework to predict when, why, and how racial groups react defensively toward others. America's racial groups can be arrayed along two dimensions: how American and how superior are they considered? This Element claims that location along these axes motivates political reactions to outgroups. Using original survey data and experiments, this Element reveals the acute sensitivity that people of color have to their social station and how it animates political responses to racial diversity.

scholarly journals Racial inequalities in the socioeconomic, demographic and health conditions of elderly from Maranhão State, Legal Amazon, Brazil: a population-based study

2014 ◽  
Vol 44 (3) ◽  
pp. 335-344 ◽  
Author(s):  
Bruno Luciano Carneiro Alves de Oliveira ◽  
Alécia Maria da Silva ◽  
Raimundo Antonio da Silva ◽  
Erika Barbara Abreu Fonseca Thomaz
Keyword(s):  
Social Role ◽  
Health Indicators ◽  
The Elderly ◽  
Population Based ◽  
Sample Survey ◽  
Health Conditions ◽  
Activity Levels ◽  
Racial Groups ◽  
Cross Sectional ◽  
Racial Inequalities

Aging with quality of life does not occur equally among the racial groups of Brazilian elderly, and few studies have analyzed this issue in the states of the Brazilian Legal Amazon. The objective of this study was to investigate racial inequalities in the socioeconomic, demographic and health conditions of elderly residents of Maranhão state, Brazil. The present work is a cross-sectional study of 450 elders aged 60 years or older included in the 2008 National Household Sample Survey. The prevalence of socioeconomic, demographic, health and habit indicators and of risk factors were estimated in white, brown and black racial categories that were self-reported by the survey participants. The chi-square test was used for comparisons (a=5%). The majority of the elderly respondents identified themselves as brown (66.4%) or white (23.3%). There were significant socioeconomic, demographic, habit and lifestyle differences among the racial groups. Most of the black and brown elderly lived alone, reported lower educational levels and were in the lowest quintile for income. These respondents were also highly dependent on the Unified Health System (Sistema Único de Saúde - SUS), exhibited low rates of screening mammograms and lower physical activity levels and had a greater proportion of smokers. However, there was no difference in the prevalence of health indicators or in the proportion of elderly by gender, age, social role in the family or the urban-rural location of the household. These results indicate the presence of racial inequalities in the socioeconomic and demographic status and in the practice of healthy habits and lifestyles among elderly from Maranhão, but suggest equity in health status. The results also suggest the complexity and challenges of interlinking race with socioeconomic aspects, and the findings reinforce the need for the implementation of public policies for these population groups.

scholarly journals Judging good appearance in personnel selection

2018 ◽  
Vol 25 (87) ◽  
pp. 656-675
Author(s):  
Altair dos Santos Paim ◽  
Marcos Emanoel Pereira
Keyword(s):  
Labor Market ◽  
Human Resources ◽  
Random Sample ◽  
Personnel Selection ◽  
Racial Bias ◽  
Selection Process ◽  
Physical Appearance ◽  
Racial Diversity ◽  
Job Applicants ◽  
Selection Of

ABSTRACT Judgement of what one views as good appearance in the selection of job applicants may reveal racial bias in access to the labor market. The purpose of this study is to evaluate the effects of racism in judging physical appearance in personnel selection. The non-random sample was composed of seventy-four (74) participants, of whom forty-two were human resources professionals (57%). The instruments used were an assessment of résumés, a set of prejudice scales, an inventory of racism in the labor market, an indicator of good appearance and a sociodemographic questionnaire. Three hypotheses were tested. Hypothesis 1, which postulated a preference for white candidates was confirmed. Hypothesis 2 was corroborated, because the professionals showed a higher tendency to choose candidates with a fairer complexion. Hypothesis 3, which made reference to good appearance was rejected, because the participants elected hygiene as a further element present in the judgment in selecting candidates. Finally, it is considered that the selection process should be based on the acceptance of racial diversity, a key element for the development of creative and innovative organizations.

Examining the Relationships Between Diversity and Work Behaviors in U.S. Federal Agencies: Does Inclusive Management Make a Difference?

2016 ◽  
Vol 38 (2) ◽  
pp. 218-247 ◽  
Author(s):  
Kuk-Kyoung Moon
Keyword(s):  
Positive Relationship ◽  
Gender Diversity ◽  
Racial Diversity ◽  
Critical Issue ◽  
Federal Agencies ◽  
Work Behaviors ◽  
Demographic Diversity ◽  
Work Related ◽  
Managing Diversity ◽  
Turnover Behavior

Workforce diversity has been depicted as a double-edged sword that leads to both positive and negative work-related outcomes. As a result, the critical issue in diversity research is concerned with enhancing the benefits and reducing the detriments of heterogeneity within organizations on work behaviors. By combining theories on diversity and inclusiveness, this article examines inclusive management at the federal subagency level as a moderator of the relationships between demographic diversity (gender and race) and work behaviors (innovative and turnover behavior). Using survey and personnel data drawn from federal subagencies, inclusive management—a set of policies aimed at recognizing all employees as valued organizational insiders with unique identities—not only strengthens the positive relationship between racial diversity and innovative behavior but also attenuates the positive relationship between gender diversity and turnover behavior. These findings suggest that inclusive management is a key strategy for effectively managing diversity.

Regulation and ethics in artificial intelligence and machine learning technologies: Where are we now? Who is responsible? Can the information professional play a role?

2020 ◽  
Vol 37 (2) ◽  
pp. 60-68
Author(s):  
Denise Carter
Keyword(s):  
Artificial Intelligence ◽  
Machine Learning ◽  
White Paper ◽  
Learning Technologies ◽  
Public And Private ◽  
General Data Protection Regulation ◽  
Oecd Principles ◽  
Information Professional ◽  
Uk Government ◽  
The Uk

Artificial intelligence (AI) and machine learning (ML) technologies are rapidly maturing and proliferating through all public and private sectors. The potential for these technologies to do good and to help us in our everyday lives is immense. But there is a risk that unless managed and controlled AI can also cause us harm. Questions about regulation, what form it takes and who is responsible for governance are only just beginning to be answered. In May 2019, 42 countries came together to support a global governance framework for AI. The Organisation for Economic Co-operation and Development (OECD) Principles on Artificial Intelligence (OECD (2019) OECD principles on AI. Available at: https://www.oecd.org/going-digital/ai/principles/ (accessed 2 March 2020)) saw like-minded democracies of the world commit to common AI values of trust and respect. In Europe, the European Commission’s (EC) new president, Ursula von der Leyen has made calls for a General Data Protection Regulation style. As a first step the EC has published a white paper: ‘On Artificial Intelligence – A European Approach to Excellence and Trust’ (European Commission (2020) Report, Europa, February). In February 2020, the UK government has published a report on ‘Artificial Intelligence in the Public Sector’ (The Committee on Standards in Public Life (2020) Artificial intelligence and public standards. Report, UK Government, February). This article discusses some of the potential threats AI may hold if left unregulated. It provides a brief overview of the regulatory activities for AI worldwide, and in more detail the current UK AI regulatory landscape. Finally, the article looks at the role that the information professional might play in AI and ML.

scholarly journals Beliefs about social norms and racial inequalities predict variation in the early development of racial bias

2021 ◽  
Author(s):  
Michael T. Rizzo ◽  
Emily R. Green ◽  
Yarrow Dunham ◽  
Emile Bruneau ◽  
Marjorie Rhodes
Keyword(s):  
Social Norms ◽  
Early Development ◽  
Racial Bias ◽  
Racial Inequalities

scholarly journals The Ethnic Dimensions in Social Movements

2017 ◽  
Author(s):  
Pamela Oliver
Keyword(s):  
Social Networks ◽  
Social Movements ◽  
Social Construction ◽  
Race And Ethnicity ◽  
Racial Bias ◽  
Group Formation ◽  
The United States ◽  
Racial Groups ◽  
Ethnic Boundaries ◽  
The Social

This paper draws on work in the social construction of race and ethnicity to explain why race/ethnic divisions are so often axes of domination and why these divisions are central to social movements. (1) Ethnic/racial groups are constructed in political processes that are tied to state formation and social movements. Many states (including the United States) have an ethnic/racial bias or footprint in their construction. Ethnic/racial groups that are numerical majorities have an advantage in determining state policies and state actions that advantage dominant groups over subordinate groups, create chains of interrelations that amplify differences in power and privilege, and take actions to prohibit or prevent reparations or redress for these past actions. (2) Network isolation and intergenerational transmission interact with structures of domination to reproduce domination over time. “Ethnicity” matters when ethnic boundaries are relatively sharp, consequential, and highly correlated with domination structures and social networks. Strong “ethnic” boundaries tend to divide societies into majorities and minorities. (3) Dominant groups develop and reproduce cultures of domination that include both hostile and benign paternalistic relations with other groups. Subordinate groups develop and reproduce cultures that intermingle opposition and submission. Collective identities are both imposed from without by the actions of others and asserted from within. Identities and cultural practices are developed collectively within social networks and influenced by the actions and speech of political actors, including social movements. (4) Regardless of whether their goals are group-oriented or issue-oriented, all movements in an ethnically-divided or ethnically-hierarchical society have an “ethnic” dimension in the sense that they draw from or map onto one or more ethnic groups. Movements arising from privileged “ethnic” majorities have different dynamics from movements by disadvantaged “ethnic” minorities or mixed-ethnic movements. Processes of group formation derived from theories of the social construction of ethnicity illuminate other movement-relevant group formation processes, including class formation and political subcultures. Lying at the intersection of the sociology of social movements and the sociology of race and ethnicity, the “ethnic” dimensions are revealed as a lens for understanding the general problems of group and identity formation and collective mobilization that lie at the heart of both areas.Presented at the 2016 meeting of the American Sociological Association. NOTE: The uploaded version is now a preprint of the 2017 published version, which is a substantial revision of the 2016 ASA version.

Detecting racial bias in algorithms and machine learning

2018 ◽  
Vol 16 (3) ◽  
pp. 252-260 ◽  
Author(s):  
Nicol Turner Lee
Keyword(s):  
Emerging Technologies ◽  
Racial Bias ◽  
Public Policies ◽  
Social Implications ◽  
Workplace Diversity ◽  
High Tech ◽  
Racial Groups ◽  
Content Type ◽  
The Government ◽  
Practical Implications

Purpose The online economy has not resolved the issue of racial bias in its applications. While algorithms are procedures that facilitate automated decision-making, or a sequence of unambiguous instructions, bias is a byproduct of these computations, bringing harm to historically disadvantaged populations. This paper argues that algorithmic biases explicitly and implicitly harm racial groups and lead to forms of discrimination. Relying upon sociological and technical research, the paper offers commentary on the need for more workplace diversity within high-tech industries and public policies that can detect or reduce the likelihood of racial bias in algorithmic design and execution. Design/methodology/approach The paper shares examples in the US where algorithmic biases have been reported and the strategies for explaining and addressing them. Findings The findings of the paper suggest that explicit racial bias in algorithms can be mitigated by existing laws, including those governing housing, employment, and the extension of credit. Implicit, or unconscious, biases are harder to redress without more diverse workplaces and public policies that have an approach to bias detection and mitigation. Research limitations/implications The major implication of this research is that further research needs to be done. Increasing the scholarly research in this area will be a major contribution in understanding how emerging technologies are creating disparate and unfair treatment for certain populations. Practical implications The practical implications of the work point to areas within industries and the government that can tackle the question of algorithmic bias, fairness and accountability, especially African-Americans. Social implications The social implications are that emerging technologies are not devoid of societal influences that constantly define positions of power, values, and norms. Originality/value The paper joins a scarcity of existing research, especially in the area that intersects race and algorithmic development.

Sign in / Sign up

Export Citation Format

Share Document