Astra: AI Safety, Trust, & Risk Assessment

TL;DR

The paper addresses the mismatch between global AI safety frameworks and India's distinct socio-technical reality, proposing ASTRA and the AI Safety Risk Database (AIRD) to ground safety in local context. It adopts a bottom-up, grounded-theory methodology to produce a domain-agnostic yet India-specific ontology with 37 leaf-leveI risk classes organized into Social Risks and Frontier/Socio-Structural Risks, focusing initially on Education and Financial Lending. The AIRD links risks to Timing, Stakeholder, and Intent, enabling domain auditors and policymakers to map and mitigate design-flaw–driven harms across India's Digital Public Goods landscape. By treating risk governance as a living utility—open to domain expansion, regulatory mapping, and community curation—the framework aims to enhance sovereignty, inclusivity, and resilience in India's AI ecosystem, with practical implications for regulators, developers, and public agencies.

Abstract

This paper argues that existing global AI safety frameworks exhibit contextual blindness towards India's unique socio-technical landscape. With a population of 1.5 billion and a massive informal economy, India's AI integration faces specific challenges such as caste-based discrimination, linguistic exclusion of vernacular speakers, and infrastructure failures in low-connectivity rural zones, that are frequently overlooked by Western, market-centric narratives. We introduce ASTRA, an empirically grounded AI Safety Risk Database designed to categorize risks through a bottom-up, inductive process. Unlike general taxonomies, ASTRA defines AI Safety Risks specifically as hazards stemming from design flaws such as skewed training sets or lack of guardrails that can be mitigated through technical iteration or architectural changes. This framework employs a tripartite causal taxonomy to evaluate risks based on their implementation timing (development, deployment, or usage), the responsible entity (the system or the user), and the nature of the intent (unintentional vs. intentional). Central to the research is a domain-agnostic ontology that organizes 37 leaf-level risk classes into two primary meta-categories: Social Risks and Frontier/Socio-Structural Risks. By focusing initial efforts on the Education and Financial Lending sectors, the paper establishes a scalable foundation for a "living" regulatory utility intended to evolve alongside India's expanding AI ecosystem.

Figures (1)

• Figure 1: Domain agnostic AI safety ontology