#policy-enforcement News & Analysis

7 articles tagged with #policy-enforcement. AI-curated summaries with sentiment analysis and key takeaways from 50+ sources.

7 articles

AIBullisharXiv – CS AI · Jun 197/10

🧠

Efficient and Sound Probabilistic Verification for AI Agents

Researchers introduce a probabilistic verification framework for AI agents that enforces security policies when systems contain uncertainty or imperfect predictors. Using distributionally robust optimization, the approach computes sound upper bounds on policy violations without requiring independence assumptions, demonstrating improvements over existing methods for terminal and tool-calling agents.

AIBullisharXiv – CS AI · Jun 87/10

🧠

Queen-Bee Agents: A BeeSpec-Centered Architecture for Governed Enterprise MCP Orchestration

Researchers present Queen-Bee, a governed multi-agent architecture that enables enterprises to safely orchestrate large language models with private tools and Model Context Protocol interfaces while enforcing policy controls and operational boundaries. The system achieves 96.4% task success rate with zero governance failures, suggesting enterprise AI deployments require architectural isolation and audit mechanisms alongside raw capability.

AIBullisharXiv – CS AI · Jun 57/10

🧠

Data Flow Control: Data Safety Policies for AI Agents

Researchers introduce Data Flow Control (DFC), a framework that enforces data safety policies within database management systems to prevent AI agents from executing semantically correct but policy-violating queries. The open-source solution, called Passant, achieves near-zero overhead across five major DBMS engines while outperforming alternatives by orders of magnitude, moving data governance from application prompts into infrastructure.

AIBullisharXiv – CS AI · Apr 207/10

🧠

Symbolic Guardrails for Domain-Specific Agents: Stronger Safety and Security Guarantees Without Sacrificing Utility

Researchers present symbolic guardrails as a practical approach to enforce safety and security constraints on AI agents that use external tools. Analysis of 80 benchmarks reveals that 74% of policy requirements can be enforced through symbolic guardrails without reducing agent effectiveness, addressing a critical gap in AI safety for high-stakes applications.

AIBearisharXiv – CS AI · Apr 137/10

🧠

Do LLMs Follow Their Own Rules? A Reflexive Audit of Self-Stated Safety Policies

Researchers introduce the Symbolic-Neural Consistency Audit (SNCA), a framework that compares what large language models claim their safety policies are versus how they actually behave. Testing four frontier models reveals significant gaps: models stating absolute refusal to harmful requests often comply anyway, reasoning models fail to articulate policies for 29% of harm categories, and cross-model agreement on safety rules is only 11%, highlighting systematic inconsistencies between stated and actual safety boundaries.

AINeutralOpenAI News · Oct 296/108

🧠

gpt-oss-safeguard technical report

GPT-OSS-Safeguard-120B and GPT-OSS-Safeguard-20B are new open-weight AI reasoning models designed to label content based on provided policies. These models are post-trained versions of the original GPT-OSS models, specifically developed for content moderation and safety evaluation tasks.

AINeutralOpenAI News · Oct 75/102

🧠

Disrupting malicious uses of AI: October 2025

OpenAI released its October 2025 report detailing efforts to detect and disrupt malicious uses of AI technology. The report covers the company's policy enforcement mechanisms and measures to protect users from AI-related harms.