- 代码
NRS-T1- 标题
- 人工智能风险
- 版本
- not yet versioned
- 状态
- 打开
- 日期
- circulated 17 October 2026
- Unique Nexus Identifier
- allocated when drafting starts (NS1-R2)
- 委员会
- TC-TEC
- 波次
- 3
- 来源
standards/risk/NRS-T1.mdat52ade713b78a
Preparatory working draft, circulated for review and comment. Not for implementation. Prepared and circulated by The Secretariat of the Nexus Standards, Global Centre for Risk and Innovation (GCRI Canada) (the contact form), 17 October 2026. Technical review area: TR-08. This document is not a Nexus Standard; it may change without notice and is not to be implemented, relied upon or cited as a Nexus Standard. Its stage is designated only by the Global Standards Council and its technical committees under NS-13. Recipients are invited to submit comments with supporting reasons, and to notify any patent rights of which they are aware. See the notice of circulation.
- 系列
- T
- 指数型技术
- 状态
- 待定(已在目录中命名并明确规定,尚未起草)
- 技术委员会
- TC-TEC(指数级技术)
- 波次
- 3(技术、金融、政策与发展)
引用
NRS-T1, clause 1. NRS-T1. https://nexustandards.com/standards/nrs/nrs-t1/#nx-c-11. 范围
- Technology stream: ai
- 领域:
AIS人工智能与模型,DRG数字与人工智能监管 - Functional system classes: NXS-NXQ (software, models and capability lifecycle), NXS-LAB (research, experiments and simulations), NXS-NSF (trust, authorisation and assurance)
- Ecosystem areas governed: Artificial intelligence (GCRI Canada field gateway; the Global Risks Forum (GRF) and the Global Risks Alliance (GRA) counterparts) (
field-artificial-intelligence);AI 架构、AI 安全评估和 人工监督、模型治理、可验证计算(ai-ml-governance); Agency (technical assistance and expert matching) (nexus-agency) - Programme areas:
sovereign-ai
引用
NRS-T1, clause 2. NRS-T1. https://nexustandards.com/standards/nrs/nrs-t1/#nx-c-22. 各部分
| 部分 | 标题 |
|---|---|
| NRS-T1-1 | 词汇表与系统概况 |
| NRS-T1-2 | 该技术的危害、失效模式和损害 |
| NRS-T1-3 | 风险应用中使用的保证:验证、不确定性与局限 |
| NRS-T1-4 | 数据、模型与供应链完整性 |
| NRS-T1-5 | 安防、安全与韧性控制措施 |
| NRS-T1-6 | 监测、事件与变更 |
| NRS-T1-7 | 治理、权利与问责 |
| NRS-T1-8 | 评估与披露报告 |
引用
NRS-T1, clause 3. NRS-T1. https://nexustandards.com/standards/nrs/nrs-t1/#nx-c-33. Nexus 原生来源
Everything this standard classifies comes from these Nexus authorities; it adds no list of its own.
universe.technology前沿技术登记册与参考基线nrm#cyber-ai-frontier-techNexus-native risk methods (Risk Master 23 fabrics, 670 capabilities, 169 bindings, 147 models, VVUQ, risk product contract, claim registry) (itemcyber-ai-frontier-tech)nhs-1#tech.ai-failureNHS-1 Nexus Hazard Standard (counts in the native register) (itemtech.ai-failure)nhs-1#tech.ai-securityNHS-1 Nexus Hazard Standard (counts in the native register) (itemtech.ai-security)nhs-1#tech.ai-misuseNHS-1 Nexus Hazard Standard (counts in the native register) (itemtech.ai-misuse)nhs-1#tech.ai-agenticNHS-1 Nexus Hazard Standard (counts in the native register) (itemtech.ai-agentic)nhs-1#tech.ai-societyNHS-1 Nexus Hazard Standard (counts in the native register) (itemtech.ai-society)nhs-1#techNHS-1 Nexus Hazard Standard (counts in the native register) (itemtech)nis-1#TECNIS-1 Nexus Indicator Standard(20 个领域,451 项指标,106 个综合指数)(条目TEC)
引用
NRS-T1, Crosswalk to external standards. NRS-T1. https://nexustandards.com/standards/nrs/nrs-t1/#nx-c-crosswalk-to-external-standards与外部标准的对照
Held at arm’s length in the external standards register and mapped element by element (NS-14); never the basis of a Nexus classification and never reproduced.
| 标准 | 机构 | 关系 | 访问 |
|---|---|---|---|
| ISO/IEC 42001:2023 Artificial intelligence – Management system | ISO/IEC JTC 1/SC 42 | 对照 | 付费墙后,尚未核实 |
| ISO/IEC 23894:2023 Information technology – Artificial intelligence – Guidance on risk management | ISO/IEC JTC 1/SC 42 | 兼容性配置文件(受限) | 付费访问 |
| ISO/IEC 42005:2025 人工智能系统影响评估 | ISO/IEC JTC 1/SC 42 | 引用 | 付费墙后,尚未核实 |
| NIST 人工智能风险管理框架(AI RMF 1.0,NIST AI 100-1) | NIST | 对照 | 开放 |
| NIST AI 600-1 人工智能风险管理框架:生成式人工智能概况 | NIST | 引用 | 开放 |
| 第 2024/1689 号条例(欧盟)(《人工智能法》) | 欧洲议会和欧盟理事会 | 对照 | 开放 |
| 联合国教科文组织《人工智能伦理问题建议书》 | UNESCO | 对照 | 开放,尚未核实 |
基础:每项 Nexus Risk Standard 均适用 NS-1(标识符)、NS-2(引用)、NS-3 和 NS-3b(记录与元数据)、NS-10(条款)、NS-11(验证)、NS-13(流程)以及 NS-14(符合性与对照)。
引用
NRS-T1, clause 4. NRS-T1. https://nexustandards.com/standards/nrs/nrs-t1/#nx-c-44. 依赖于
引用
NRS-T1, clause 5. NRS-T1. https://nexustandards.com/standards/nrs/nrs-t1/#nx-c-55. 符合性
| 一致性等级 | 条款 | 含义 |
|---|---|---|
| CL-Base | NXOSI 8.4.1.1 | 制品模式、回执格式和状态语义满足基线要求;验证库产生一致的结果;基本配置文件的导入/导出在声明的约束下可用。 |
| CL-Operational | NXOSI 8.4.1.2 | Runtime enforcement points, docketing, control-plane auditability, and clock enforcement are implemented with defined degraded-mode behaviour and traceable artefacts. |
| CL-Assured | NXOSI 8.4.1.3 | 已对规定类别实施证明优先的要求;评估策略得到执行;状态和更正工作流证明在变化中依赖关系保持稳定。 |
| CL-Sovereign | NXOSI 8.4.1.4 | 主权区域执行、数据就地计算模式、离线核验包和受控披露工作流均已实施,并附可供审计的访问日志。 |
| CL-Federated | NXOSI 8.4.1.5 | Multi-organisation federation is demonstrated: boundary enforcement, cross-org routing constraints, interoperable ontology exchange, and consistent verification across parties. |
| 证据质量等级 | 条款 | 含义 |
|---|---|---|
| EQL-1 | NXOSI 8.4.2.1 | 证据仅限于可公开的安全摘要或信息性摘要,以及适合低风险依赖的回执级工件。 |
| EQL-2 | NXOSI 8.4.2.2 | 检查可重复并与遥测数据关联;证据足以支持运营补救和内部保证。 |
| EQL-3 | NXOSI 8.4.2.3 | 证据包包含足够的原始工件、来源追溯和关联,可在受控披露下供审计依赖。 |
| EQL-4 | NXOSI 8.4.2.4 | 高后果证据包括已测量的状态绑定以及在已声明输入下可重放的核查,从而能够形成高度依赖。 |
| EQL-5 | NXOSI 8.4.2.5 | 证据通过受控的披露工作流、严格的处理、合法依据矩阵和完整的记录纪律,为监管依赖提供支持。 |
Notation: <NRS id>:<conformance level>/<evidence quality level>, for example NRS-N1:CL-Assured/EQL-4. Source: NXOSI part VIII section 8.4, with the NRM evidence quality scale.
要求以机器可读条款的形式发布于 NRS-C10 库,并带有标识符 NRS-T1-<part>-R<n>,每项均附测试向量,因此 CL-Operational 与 EQL-2 及以上可由机器检查。符合性声明写明版本、部分、等级和证据质量等级,例如 NRS-T1:CL-Assured/EQL-3.
引用
NRS-T1, clause 6. NRS-T1. https://nexustandards.com/standards/nrs/nrs-t1/#nx-c-66. 起草
起草工作由 TC-TEC 负责。其交付各部分的规范性文本、条款和测试向量、对照记录、示例以及 Academy 模块(计划主线交接 F09)。全球标准理事会负责通过;各国理事会以六种采纳模式之一予以采纳或调整;此处任何内容均不构成认证或就绪性声明。
变更历史
Versions from the catalogue; entries from the changelog that name this document.
No version is recorded in the catalogue for this document yet.
更新日志中没有任何条目提及 NRS-T1。