Skip to content

GCRI · 联合体各领域

人工智能

AI risk management asks what a system can do, how it fails, who is affected and who is accountable. The lifecycle matters: capability can change after deployment, and harms can be diffuse.

02三项职责

从三个角度看该领域。

03系统性问题

From model failure to systemic effect.

Individual AI failures (biased outputs, hallucinated facts, unsafe actions by an agent) are well catalogued. Systemic effects arise when many actors use similar models trained on similar data, when AI is embedded in critical decisions such as credit or triage, or when capability shifts faster than oversight. Assessment has to cover the model, the system it sits in and the institution operating it.

图层典型风险评估
建模Robustness, bias, capability for misuseEvaluations and red teaming
系统Integration errors, over-reliance, data leakage影响评估、 人工监督 设计
机构Unclear accountability, missing incident responseManagement system audit (ISO/IEC 42001)
生态系统Homogeneity, concentration, correlated failure市场和依赖关系分析
04框架

风险框架与法律。

框架或机构其管辖或提供的内容
NIST AI RMF 及其生成式人工智能概况治理、映射、衡量和管理四项职能
ISO/IEC 42001 和 ISO/IEC 23894AI management systems and AI risk management guidance
欧盟《人工智能法》按风险分级的义务分阶段实施;请核对各项义务的现行日期
经合组织人工智能原则和联合国教科文组织《人工智能伦理问题建议书》Intergovernmental principles
05耦合

Where AI risk lands.

The science sits on AI and Autonomous Systems; methods start from AI assurance.

08问题

问题与解答。

GCRI 为何托管现场网关?
GCRI hosts the shared references of the Consortium, including the STEM Commons and the field gateways, so each field has one place that links all three institutions’ work.
报告贡献

从本页贡献证据。

Reports 是联合体的季刊。您对该主题的了解可以成为下一期中的证据、案例或更正,署名归您,并在属于您本人作品时以 CC BY 4.0 发布。

GCRI
全球风险与创新中心(The Global Centre for Risk and Innovation)

带来一种方法、一个数据集或一个问题。

信息和咨询现已开放。每项咨询都会获得一个编号,并告知下一步;申请、付款、算力分配和注册只有在相应团队、条款和审查能力就绪时才会开放。