Public Section Preview
Quick Revision Table
| Concept | Key Fact | RPSC Relevance |
|---|---|---|
| AI in administration | Pattern recognition, predictive decisions | New 2026 topic — high probability |
| Algorithmic bias | AI replicates training data discrimination | Welfare exclusion, policing examples |
| Accountability gap | "Problem of many hands" (Floridi) | Who is responsible when AI errs? |
| Black-box problem | Unexplainable AI decisions | Natural justice, Article 14 violation |
| Explainability (XAI) | AI must justify its recommendations | LIME, SHAP techniques; Digital India Act |
| HITL | Human must be final decision-maker | High-stakes admin decisions |
| Zweckrationalität | Means-end rationality (Weber) — AI's mode | Vs. value rationality of conscience |
| Wertrationalität | Value rationality — conscience's mode | Weber's distinction |
| Kant on AI | AI cannot be a Kantian moral agent | No will, no duty, no dignity |
| Phronesis (Aristotle) | Practical wisdom — hallmark of good admin | Cannot be programmed |
| Compassion deficit | AI cannot feel empathy | Widow pension, drought victim examples |
| DBT and exclusion errors | Aadhaar-DBT has genuine beneficiary exclusions | Human override needed |
| Rajasthan context | Jan Aadhar, e-Dharti, NDVI drought AI | Collector retains conscience duty |
| Responsible AI for All | NITI Aayog, 2021 | India's ethical AI framework |
| India AI Mission | Rs 10,372 crore, 2024 | Compute, data, safety focus |