Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Ethical Fairness without Demographics in Human-Centered AI

Shaily Roy, Harshit Sharma, Asif Salekin

Abstract

Computational models are increasingly embedded in human-centered domains such as healthcare, education, workplace analytics, and digital well-being, where their predictions directly influence individual outcomes and collective welfare. In such contexts, achieving high accuracy alone is insufficient; models must also act ethically and equitably across diverse populations. However, fair AI approaches that rely on demographic attributes are impractical, as such information is often unavailable, privacy-sensitive, or restricted by regulatory frameworks. Moreover, conventional parity-based fairness approaches, while aiming for equity, can inadvertently violate core ethical principles by trading off subgroup performance or stability. To address this challenge, we present Flare (Fisher-guided LAtent-subgroup learning with do-no-harm REgularization), the first demographic-agnostic framework that aligns algorithmic fairness with ethical principles through the geometry of optimization. Flare leverages Fisher Information to regularize curvature, uncovering latent disparities in model behavior without access to demographic or sensitive attributes. By integrating representation, loss, and curvature signals, it identifies hidden performance strata and adaptively refines them through collaborative but do-no-harm optimization, enhancing each subgroup's performance while preserving global stability and ethical balance. We also introduce BHE (Beneficence-Harm Avoidance-Equity), a novel metric suite that operationalizes ethical fairness evaluation beyond statistical parity. Extensive evaluations across diverse physiological (EDA), behavioral (IHS), and clinical (OhioT1DM) datasets show that Flare consistently enhances ethical fairness compared to state-of-the-art baselines.

Ethical Fairness without Demographics in Human-Centered AI

Abstract

Computational models are increasingly embedded in human-centered domains such as healthcare, education, workplace analytics, and digital well-being, where their predictions directly influence individual outcomes and collective welfare. In such contexts, achieving high accuracy alone is insufficient; models must also act ethically and equitably across diverse populations. However, fair AI approaches that rely on demographic attributes are impractical, as such information is often unavailable, privacy-sensitive, or restricted by regulatory frameworks. Moreover, conventional parity-based fairness approaches, while aiming for equity, can inadvertently violate core ethical principles by trading off subgroup performance or stability. To address this challenge, we present Flare (Fisher-guided LAtent-subgroup learning with do-no-harm REgularization), the first demographic-agnostic framework that aligns algorithmic fairness with ethical principles through the geometry of optimization. Flare leverages Fisher Information to regularize curvature, uncovering latent disparities in model behavior without access to demographic or sensitive attributes. By integrating representation, loss, and curvature signals, it identifies hidden performance strata and adaptively refines them through collaborative but do-no-harm optimization, enhancing each subgroup's performance while preserving global stability and ethical balance. We also introduce BHE (Beneficence-Harm Avoidance-Equity), a novel metric suite that operationalizes ethical fairness evaluation beyond statistical parity. Extensive evaluations across diverse physiological (EDA), behavioral (IHS), and clinical (OhioT1DM) datasets show that Flare consistently enhances ethical fairness compared to state-of-the-art baselines.
Paper Structure (55 sections, 5 equations, 5 figures, 8 tables, 2 algorithms)

This paper contains 55 sections, 5 equations, 5 figures, 8 tables, 2 algorithms.

Table of Contents

  1. Introduction
  2. Related Work
  3. Fairness through Demographic Information
  4. Fairness without Demographic (FWD)
  5. Subgroup-oriented Fairness
  6. Ethical AI in healthcare and aligning with human-centered modeling
  7. Ethical AI vs fairness
  8. Fairness Metrics
  9. Limitations of existing fairness metrics
  10. BHE: Fairness Metrics Aligned with Bioethics
  11. Design Choice
  12. Base pretraining.
  13. Latent Clustering
  14. Cluster-specific Adaptation and Aggregation
  15. Cluster Adaptation:
  16. ...and 40 more sections

Figures (5)

  • Figure 1: Proposed Approach: (1) Base pre-training learns embeddings with Fisher penalty regularization, (2) UMAP+GMM clustering captures behavior-based similarity, and (3) cluster models are fine-tuned and aggregated with stability constraints
  • Figure 2: Overall F1-scores across all models for each dataset.
  • Figure 3: User-level F1 mean and F1 standard deviation across all models for the OhioT1DM, EDA, and IHS datasets. Lower F1 std indicates more consistent model behavior, supporting user-level autonomy.
  • Figure 4: Fold and Cluster-wise F1 comparison across datasets within the proposed Flare framework. Bars show F1-scores for Base preTraining (BpT, light brown) as the intermediate module (1) and Flare (dark brown) as the final module (3) across all fold--cluster pairs. Values above each pair ($\Delta$) indicate the change from BpT to Flare.
  • Figure 5: Loss landscape visualizations for the benign baseline, base Pre-training without Fisher penalty, base pretraining with Fisher penalty regularization, and Flare across three datasets. The red, orange, and green arrows and dotted circles respectively highlight curvature transitions, showing how Fisher regularization and cluster adaptation progressively flatten the optimization landscape.