4arXiv cs.AI (Artificial Intelligence)·
TrustX Agent Risk Classification Framework (ARC) for enterprise agentic AI governance
Researchers introduce the TrustX Agent Risk Classification (ARC) Framework, a structured instrument for risk-tiering internally created agentic AI systems across seven system types. The framework centers on a twelve-dimension scoring rubric combined with a GPA+IAT classification model and a five-level autonomy scale, producing a three-tier governance output with control recommendations. A specialized extension for coding assistants is included, and the framework is publicly accessible as an interactive tool at arc.responsible.ai. The work targets AI governance practitioners, risk officers, and regulators dealing with enterprise and public-sector agentic deployments.
Related guides (3)
Related events (8)
Anthropic publishes framework for safe and trustworthy agent development
Anthropic released a formal framework for responsible agent development, articulating principles around human oversight, transparency, value alignment, and privacy for autonomous AI agents. The document draws on Claude Code as a reference implementation and cites enterprise deployments at Trellix and Block as real-world examples. The framework is positioned as a contribution to emerging industry standards for agentic AI systems, acknowledging open technical challenges in value alignment measurement and oversight calibration.
Eticas AI Risk Taxonomy v2.0.0: Open infrastructure for operationalizing AI audits
Eticas presents a structured AI auditing framework that bridges risk cataloging to executable audit methodology, demonstrated end-to-end on PII leakage testing against GPT-4-0314. The taxonomy organizes 76 active subcategories across 10 categories with mappings to 18 external frameworks, and is published under CC BY 4.0 with SKOS/JSON-LD distributions. The key contribution is an operationalization layer that converts named risks into measurable, severity-graded findings — addressing a gap the authors identify across at least 74 existing AI risk taxonomies. The PII leakage demonstration shows disclosure rates ranging from 0% to 84% under adversarial conditioning, graded as SYSTEMIC severity.
Anthropic publishes Responsible Scaling Policy with AI Safety Level framework
Anthropic released its Responsible Scaling Policy (RSP), a formal framework of technical and organizational protocols for managing catastrophic risks from increasingly capable AI systems. The policy introduces AI Safety Levels (ASL-1 through ASL-5+), modeled on US biosafety level standards, requiring progressively stricter safety, security, and operational standards as models become more capable. Current Claude models are classified as ASL-2; ASL-3 triggers stricter deployment constraints including adversarial red-teaming requirements. The policy has been approved by Anthropic's board and is intended as a template for industry-wide adoption.
Microsoft Agent Governance Toolkit: Policy Enforcement and Zero-Trust Security for Autonomous AI Agents
Microsoft has published an open-source Agent Governance Toolkit on GitHub covering policy enforcement, zero-trust identity, execution sandboxing, and reliability engineering for autonomous AI agents. The toolkit claims full coverage of the OWASP Agentic Top 10 security risks. It has accumulated 1,828 stars with 113 added today, indicating active community interest. This positions Microsoft as a contributor to emerging standards for safe agentic AI deployment.
Guard Rail Validation framework for runtime validation of AI agent decisions in autonomous telecom networks
A new arXiv preprint proposes the Guard Rail Validation (GRV) framework, a runtime architecture for intercepting and validating AI-driven decisions before they execute in autonomous telecommunications networks (Levels 4-5). The framework scores decisions across dimensions including action scope, reversibility, and service criticality, then applies graduated validation mechanisms ranging from logging to multi-agent consensus. The paper also addresses cross-agent conflict detection and regulatory compliance with EU AI Act Article 14, and evaluates the framework against known AI/ML attack vectors in an O-RAN deployment model.
Systematic review lays groundwork for agentic AI governance frameworks
A new arXiv preprint presents a systematic literature review on governance of agentic AI systems, identifying features that distinguish agentic AI from traditional generative systems and why those features warrant targeted regulatory attention. The authors synthesize prevailing governance priorities, proposed mechanisms, and stakeholder roles emerging in the field. The paper positions itself as preliminary groundwork for a structured governance roadmap, framing 2025 as a pivotal year for agentic AI deployment.
RubricsTree: Scalable hierarchical rubric framework for evaluating personal health AI agents
RubricsTree is a new evaluation framework for LLM-powered personal health agents, built around a hierarchical taxonomy of over 100 clinically-verifiable Boolean rubrics derived from 4,000 real user queries and curated with physician oversight. A context-aware router activates only relevant rubrics per query, enabling scalable yet expert-aligned evaluation. The framework outperforms strong LLM-as-a-judge baselines on expert alignment and, when used as training signal, yields up to ~66% relative gains on HealthBench across Gemini, GPT, and Qwen model families. The work addresses a concrete bottleneck in clinical deployment of health AI: the cost-quality tradeoff in evaluation.
Action-graded severity scale for red-teaming tool-using AI agents beyond binary attack-success rate
A new arXiv paper argues that binary attack-success rate metrics for agentic red-teaming discard critical defender-relevant information about harm severity. The authors introduce a seven-level ordinal harm rubric (L0–L6) grading an agent's tool-call trajectory by reversibility, scope crossing, and privilege escalation, computed via both a deterministic oracle and a three-model LLM judge panel. Applied to four victim models and two defenses on the AgentDojo benchmark suite, the rubric exposes cases the binary metric misses—including a defense reporting zero attack-success rate that still permits cross-scope data leaks. The judge panel achieves high ordinal agreement (Krippendorff's alpha = 0.91) but shares systematic blind spots around escalation chain recognition.