Adnan KhanAdnanProduct & AI
San Francisco Bay Area
Get in touch

Source note3 min read

The Eticas AI Risk Taxonomy: Open Infrastructure for Operationalizing AI Audits

Source

arXiv

arxiv.org/abs/2607.02201v1

Read the source

Source

Plain-Language Brief

The Eticas AI Risk Taxonomy v2.0.0 is an open-source framework for operationalizing AI audits. It provides a standardized method for evaluating AI systems across various high-stakes domains, focusing on turning risks into testable and measurable values. The taxonomy includes 76 active subcategories organized across 10 categories and 20 sub-groups, with mappings to 18 external frameworks. The authors demonstrate the use of this taxonomy by measuring disclosure risk on GPT-4-0314 and providing a worked subcategory example.

What the Document Says

  • The Eticas AI Risk Taxonomy v2.0.0 is presented as an operationalization layer for AI audits.
  • It organizes 76 active subcategories across 10 categories and 20 sub-groups, with mappings to 18 external frameworks.
  • The taxonomy provides a method for measuring risks, such as disclosure risk on GPT-4-0314, by turning them into testable and measurable values.
  • The authors demonstrate the use of this taxonomy by measuring disclosure risk at 0%, 51%, and 84% on GPT-4-0314 under increasing adversarial conditions.

Key Concepts Mentioned

  • AI Risk Taxonomy
  • Eticas AI Risk Taxonomy v2.0.0
  • Operationalization layer
  • AI audits
  • High-stakes domains
  • Competing risk taxonomies
  • PII leakage
  • GPT-4-0314
  • Disclosure risk
  • Adversarial conditioning
  • Calibrated severity bands
  • Subcategory grade
  • SYSTEMIC pattern
  • CC BY 4.0
  • Open semantic infrastructure
  • SKOS/JSON-LD distributions
  • AI auditing field

Claims Made by the Document

  • The Eticas AI Risk Taxonomy v2.0.0 is a useful tool for operationalizing AI audits.
  • The taxonomy provides a standardized method for measuring risks across various domains.

Evidence Used

  • The paper describes the Eticas AI Risk Taxonomy and its use in measuring disclosure risk on GPT-4-0314.
  • The authors demonstrate the operational layer of the taxonomy through a worked subcategory example.

Methods / Approach

UNKNOWN

Limitations Stated by the Authors

UNKNOWN

Limitations Noted by the Agent

Agent interpretation — not from the source.

UNKNOWN

Open Questions

UNKNOWN

Notes for Future Q&A

May include interpretation. Labeled where it does.

UNKNOWN

Raw Abstract / Excerpt

The rapid deployment of AI systems across high-stakes domains has created urgent demand for standardized evaluation, yet the field remains fragmented across competing risk taxonomies that catalog risks without showing how an audit is executed. At least 74 AI risk taxonomies exist, and almost all stop at the catalog. The hard part of auditing is not naming a risk but operationalizing it: turning it into a test run against a real system, a measured value, a calibrated severity, and a defensible grade. This paper leads with that bridge. We present the operationalization layer Eticas has built and run, shown end to end on a single risk (PII leakage) against a public benchmark, and then the open taxonomy that makes the method scale. On GPT-4-0314, a disclosure risk that seven external frameworks require be controlled is measured at 0%, 51%, and 84% disclosure as adversarial conditioning increases, mapping through calibrated severity bands to a subcategory grade of E with a SYSTEMIC pattern. Around this example, the Eticas AI Risk Taxonomy v2.0.0 organizes 76 active subcategories across 10 categories and 20 sub-groups, with mappings to 18 external frameworks across compliance, reference, and academic tiers. Its category and sub-group layer is published under CC BY 4.0 as open semantic infrastructure with stable URIs and SKOS/JSON-LD distributions, and a worked subcategory example shows the operational layer down to its severity thresholds. The contribution is the demonstrated bridge from concept to graded finding, anchored by a clean separation of risks from the mechanisms by which they surface, and framed by an open-core model in which the conceptual scaffold is open and the methodology calibration is the practitioner layer. This is the infrastructure the AI auditing field needs: shared, open, and demonstrably operable.