Discriminatory Risk Assessment

A Discriminatory Risk Assessment is a risk evaluation process that incorporates protected characteristics or proxy variables as risk factors, resulting in systematically biased risk scores that disadvantage specific population groups.

• AKA: Biased Risk Evaluation, Prejudicial Risk Analysis, Discriminatory Risk Scoring.
• Context:
  • It can typically incorporate Demographic Factors into discriminatory risk assessment models.
  • It can typically use Geographic Proxys encoding discriminatory risk assessment racial bias.
  • It can typically produce Risk Scores reflecting discriminatory risk assessment prejudice.
  • It can typically influence Decision-Making Processes through discriminatory risk assessment recommendations.
  • It can typically appear Objective while embedding discriminatory risk assessment subjective bias.
  • ...
  • It can often utilize Historical Data perpetuating discriminatory risk assessment past discrimination.
  • It can often employ Machine Learning Algorithms amplifying discriminatory risk assessment bias patterns.
  • It can often resist Audit through discriminatory risk assessment technical complexity.
  • It can often create Feedback Loops reinforcing discriminatory risk assessment inequality.
  • ...
  • It can range from being a Manual Discriminatory Risk Assessment to being an Automated Discriminatory Risk Assessment, depending on its discriminatory risk assessment technology level.
  • It can range from being a Transparent Discriminatory Risk Assessment to being an Opaque Discriminatory Risk Assessment, depending on its discriminatory risk assessment interpretability.
  • It can range from being a Single-Factor Discriminatory Risk Assessment to being a Multi-Factor Discriminatory Risk Assessment, depending on its discriminatory risk assessment complexity.
  • ...
  • It can be implemented in Financial Systems determining discriminatory risk assessment credit decisions.
  • It can be applied in Criminal Justice Systems affecting discriminatory risk assessment sentencing outcomes.
  • It can be used in Healthcare Systems influencing discriminatory risk assessment treatment allocations.
  • It can be deployed in Employment Systems guiding discriminatory risk assessment hiring decisions.
  • ...
• Example(s):
  • Discriminatory Risk Assessment Tool Types, such as:
    • Geographic Discriminatory Risk Assessments, such as:
      • Neighborhood Racial Investment Risk Map grading areas by racial composition.
      • Zip Code Risk Score penalizing minority-populated areas.
      • Redlining Algorithm identifying high-risk neighborhoods.
    • Predictive Discriminatory Risk Assessments, such as:
      • Recidivism Risk Algorithm overestimating minority defendant risk.
      • Health Risk Score undervaluing minority patient needs.
      • Insurance Risk Model overcharging protected class members.
    • Credit Discriminatory Risk Assessments, such as:
      • Alternative Data Credit Score using discriminatory proxy variables.
      • Thin File Risk Assessment penalizing immigrant applicants.
  • Discriminatory Risk Assessment Implementation Domains, such as:
    • Financial Discriminatory Risk Assessments in lending decisions.
    • Insurance Discriminatory Risk Assessments in coverage determinations.
    • Employment Discriminatory Risk Assessments in candidate screening.
  • Discriminatory Risk Assessment Historical Evolutions, such as:
    • Actuarial Discriminatory Risk Assessments using statistical discrimination.
    • Algorithmic Discriminatory Risk Assessments employing machine learning bias.
  • ...
• Counter-Example(s):
  • Fair Risk Assessment, which excludes protected characteristics and discriminatory proxys.
  • Blind Risk Evaluation, which anonymizes applicant identity to prevent bias.
  • Equitable Scoring Model, which corrects for historical discrimination rather than perpetuating it.
  • Evidence-Based Risk Assessment, which uses only relevant risk factors rather than discriminatory criteria.
• See: Risk Evaluation Process, Discriminatory Financial Practice, Systemic Discrimination Mechanism, Algorithmic Bias, Statistical Discrimination, Fair Lending.

---