arXiv:2608.23870cs.AI2026-08

为生成式AI定制可共享的安全策略,支持从训练到运行的全流程管控。

Granite.Trust Policy Tools: Shareable, Actionable Policies for Generative AI Applications

论文配图:Granite.Trust Policy Tools: Shareable, Actionable Policies for Generative AI Applications
图 1 · 摘自论文原文
  • 用YAML格式定义内容约束策略,支持例外追踪与灵活治理。
  • 构建合成数据流水线,生成符合政策的训练与测试数据。
  • 开源工具链助力组织快速落地个性化安全策略。

生成式AI的安全策略需因组织、场景、法规和用户角色而异。现有访问控制方法无法有效捕捉生成式AI的内容约束需求。本文提出两项贡献:(1) 可操作策略(Actionable Policy)模式,一种基于YAML的格式,用于明确模型响应允许或禁止的内容,并支持例外管理以追踪违规;(2) 合成数据生成流水线,用于生成符合政策的训练与测试数据,并配套工具帮助策略定义与执行。二者协同支持组织一次性制定策略,并贯穿模型对齐、部署与运行监控全生命周期。策略模板、示例及工具已开源:https://github.com/ibm-granite/granite.trust.policy-tools,欢迎反馈与贡献。

原文摘要 · Abstract (English)

When it comes to safety policies for generative AI, one size does not fit all. Each organization and use case needs to mitigate different risks depending on the application context, regulatory environment, organizational values, and user personas. Yet, existing policy specification approaches are designed for traditional access control and fail to capture the nuances of GenAI application: the enforcement of content-based constraints. We present two contributions to address this gap: (1) the Actionable Policy schema, a YAML-based format for specifying what model responses can and cannot contain. The schema enables exception-based policy governance, proposing exceptions to track policy violations; (2) synthetic data generation pipeline that produces policy-aligned training data for model alignment and testing, and a set of tools to help define the schema and enforce policy. Together, these enable organizations to specify policies once and enforce them throughout the GenAI application lifecycle: from model alignment to runtime monitoring. The Actionable Policy schema, example policies, and tools are available as open source: https://github.com/ibm-granite/granite.trust.policy-tools We welcome new ideas, contributions and feedback.

生成式AI安全策略YAML开源工具

Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品;中文卡片由大模型生成,请以原文为准。