AI Ethics & Interpretability

Key concepts in responsible AI: fairness, bias detection, explainability (XAI), model cards, and regulatory compliance (EU AI Act).

AI & ML Terms
19 entries
4 sections
CC0 — public domain

The data

Core principles

Principle	Definition	Metrics	Bias sources	Mitigation	Notes	Techniques	Use cases	Practices	Regulations	Threats	Defenses
Fairness	Models should treat individuals and groups fairly, without unjust discrimination.	Demographic parity Equal opportunity Equalized odds Predictive parity	Training data bias (historical) Labeling bias Sampling bias Algorithmic bias (loss function)	Pre-processing (reweighing, disparity remover) In-processing (fairness constraints) Post-processing (threshold adjustment)	Fairness often trades off with accuracy; no universal definition — choose metrics based on context and regulation	null	null	null	null	null	null
Transparency & Explainability	Models should be interpretable; decisions explainable to stakeholders.	null	null	null	Trade-off: model complexity vs interpretability. Deep nets need post-hoc methods; linear models intrinsically explainable	Intrinsic Decision trees (inherently interpretable) Linear models (coefficients) Rule-based systems Post hoc LIME (local surrogate) SHAP (Shapley values) Partial dependence plots Feature importance	Regulated domains (finance, healthcare) Debugging model errors Building user trust	null	null	null	null
Accountability & Governance	Clear responsibility for AI system outcomes; audit trails and human oversight.	null	null	null	EU AI Act requires conformity assessments for high-risk AI systems	null	null	Model cards (metadata, intended use, limitations) Dataset documentation (datasheets) Impact assessments Human-in-the-loop for high-stakes decisions	null	null	null
Privacy & Data Protection	Respect user privacy; minimize data collection; secure training data.	null	null	null	Differential privacy provides mathematical guarantee; federated learning keeps data on device but model updates may leak info	Differential privacy (noise addition) Federated learning (on-device training) Secure multi-party computation Homomorphic encryption (research)	null	null	GDPR (EU) CCPA (California) HIPAA (health data)	null	null
Safety & Robustness	Models should be secure against adversarial attacks and fail safely.	null	null	null	Security for ML is arms race; no perfect defense against all attacks	null	null	null	null	Adversarial examples (evasion) Data poisoning (training) Model stealing/extraction Membership inference	Adversarial training Input validation Gradient masking (limited effectiveness) Certified robustness

Key concepts

Shapley values: Game theory concept; fair credit allocation among features. Computationally expensive (O(2^n)), approximated by SHAP.
Local vs global: Local: explain single prediction (LIME, SHAP). Global: overall model behavior (feature importance, partial dependence).
Monotonicity: Feature relationship with prediction should be intuitive (e.g., higher credit score → higher approval). Constrain models if needed.
Counterfactual: What minimal change would flip prediction? Useful for individual explanations and fairness analysis.

Regulations

Eu ai act

Risk levels: Unacceptable risk (banned)
High risk (strict requirements)
Limited risk (transparency)
Minimal risk (unregulated)
High risk categories: Biometric identification
Critical infrastructure
Education/employment
Public services
Law enforcement
Requirements: Risk management system
Data governance
Technical documentation
Human oversight
Conformity assessment
Status: Proposed 2021, expected 2025-2026 enforcement

Us algorithmic accountability

Approach: Sector-specific (FTC, EEOC, HHS enforce existing laws)
Key laws: Fair Credit Reporting Act (FCRA)
Equal Credit Opportunity Act (ECOA)
Civil Rights Act (Title VII)
Notes: No federal AI law yet; states passing own laws (Colorado, California)

Practical checklist

Audit training data for representation gaps and historical bias
Test model performance across demographic slices (race, gender, age)
Document model limitations and intended use cases (model card)
Implement monitoring for drift and degradation in production
Establish human review process for high-stakes decisions
Consider interpretability needs early in model selection (choose inherently interpretable if regulated)
Perform adversarial testing (worst-case inputs)
Ensure data consent and privacy compliance

Fetch the same bytes

The static files are identical to what the API returns, but with no rate limit and no server round trip. Use the API when you want a query and a content type; use the files when you want to cache one document.

curl "https://yjtoon.com/api/dataset/ai-ethics-interpretability?format=toon"

const res = await fetch(
  "https://yjtoon.com/static-data/dataset/ai-ethics-interpretability.toon"
);
const toon = await res.text();

Rate limit: 120 requests per minute per IP, no key and no signup. API reference →