arXiv AI

Just AI Governance and the Forbidden-Boundary Principle

The paper proposes a framework for evaluating ‘just AI’ across normative, data, model, and governance layers, mapping six justice requirements to evidence, decision rights, and remedies. It introduces a two‑stage evaluation rule that first excludes uses violating minimum rights constraints, then compares remaining alternatives on performance and governance criteria. The authors also present a Forbidden‑Boundary Principle and a First‑Principle Safe‑Stop Protocol, outlining policy monitoring, execution gating, safe‑state control, audit trails, and human recovery authorization, while noting that the proposal is theoretical and not yet implemented.

arXiv AI
3d ago

Between Niti and Nyaya: A Society-Centered Framework for AI Governance

The paper introduces the Society-Centered AI Governance Framework, a tool that lets policymakers, researchers, and the public evaluate AI governance instruments by assessing whether individuals can receive notice, challenge decisions, and see corrections. It applies this framework to North Carolina, revealing a visibility-first approach where notice is an agency duty but individuals lack a right to challenge or correct AI outcomes, and notes that federal preemption may keep state recourse limited. The framework aims to identify gaps and risks—such as ‘legitimacy before protection’—early in policy formation to prevent harm.

By Snehalkumar `Neil' S. Gaikwad, Punya Aragula, Shiva Kruthi Palpanoori, Pratik Patekar
arXiv AI
Oct 1

Trust Is Not a Score: Runtime Assurance Contracts for High-Risk AI Agents

The paper introduces Runtime Assurance Contracts (RAC) as a formal policy framework for high‑risk AI agents, addressing the "assurance‑transition gap" by binding autonomy boundaries, component eligibility, evidence state, transition policy, human‑review capacity, and non‑compensatory gates. RAC allows soft metrics to influence routing while mandating retries, switches, escalations, deferrals, or stops when mandatory gates fail or are unknown, ensuring aggregate performance cannot alone authorize action. The authors define the contract, evidence record, permission rule, and five invariants, and evaluate RAC through deterministic failure‑injection studies, hand‑authored traces, and a prospective synthetic holdout, comparing it to score‑only and restricted protocol baselines.

By Serhii Zabolotnii
arXiv AI
Sep 15

Governing at Machine Speed: An Adaptive Intelligence Architecture for Real-Time AI Policy Enforcement

The paper "Governing at Machine Speed: An Adaptive Intelligence Architecture for Real-Time AI Policy Enforcement" highlights a gap in enterprise AI governance, where 78% of organizations lack auditable evidence of policy enforcement. It introduces AGIL, a five-layer adaptive governance architecture that uses machine learning for real-time detection, risk classification, sub-100ms policy enforcement, continuous attestation, and policy evolution. The authors argue that the failure is organizational and architectural, not technical, and call for future empirical validation of AGIL.

By Sandeep Bokkasam, B. Durgalakshmi
arXiv AI
Sep 18

Governance-as-Code: Translating EU AI Act Technical Requirements into Executable Compliance Pipelines for Generative AI Systems

The paper introduces Governance-as-Code (GaC), a framework that translates the EU AI Act’s technical requirements into 43 machine‑checkable acceptance criteria across six compliance modules. GaC runs within a CI/CD pipeline, producing Article‑indexed audit evidence and providing actual Rego policy code. The authors validate GaC on two enterprise deployments, showing it reproduces manual audit findings—including three penalty‑triggering violations—while reducing audit labor by about 75%.

By Rudrendu Kumar Paul, Sourav Nandy
arXiv AI
Sep 2

A Mathematical Framework for Legacy, Governance, and Decision Integrity in Enterprise AI

The paper introduces a design‑science framework for ensuring legacy, governance, and decision integrity in enterprise AI systems. It defines a normalized Legacy Score based on knowledge retention, governance, oversight, adaptability, feedback learning, and jurisdictional fidelity, along with Decision Confidence and Decision Risk models, authority‑aware retrieval, Decision Memory, Regulatory Change Velocity, and a federated regulatory knowledge‑graph architecture. The authors also propose eight AI Decision Integrity Rules, an evaluation protocol, and a reproducible computational demonstration using stress tests and Monte Carlo simulations to illustrate the framework’s properties.

By Shorab Sarker