ISO/IEC 23894:2023 (AI risk mgmt)

Informational data, not legal advice. Summaries are TwinEthos's own words and rules have not been reviewed by a lawyer: check the official text before relying on any of it. A guard addresses an item; adding it is not a statement that your code meets any law.

Official text: www.iso.org.

Standard / soft law

AI-specific risk management should follow a documented identify-analyse-evaluate-treat-monitor process (ISO/IEC 23894)

ISO/IEC 23894:2023 (AI risk management process; Annexes A/B/C) · official text · Soft law or guidance (not binding law)

ISO/IEC 23894:2023 is international guidance for AI risk management. In our own words from its public scope: it guides organizations that develop, produce, deploy, or use AI-enabled products, systems, and services in managing AI-specific risk and integrating that risk management into their existing activities and functions. It adapts ISO 31000:2018 concepts to AI-related objectives, risk sources, and life-cycle activities, including post-deployment as well as design. It can complement an ISO/IEC 42001 AI management system but is not itself certifiable. Detect an AI system with no documented AI-specific risk-management process for identifying, analysing, evaluating, treating, and monitoring risks across the life cycle. [TIER C: own-words summary of public scope; licensed normative text not stored.]

Who it applies to

Duty falls on: developer, deployer
Systems covered: automated decision
Organizations that develop, produce, deploy, or use AI systems. Voluntary GUIDANCE (not certifiable — do not treat 'ISO 23894 certification' claims as valid). Used as the methodological substrate for the risk-assessment clauses of ISO/IEC 42001 and to produce documentation the EU AI Act and NIST AI RMF expect.
Whether it applies depends on facts outside the code; a person has to decide.

The guard to add

Organizational artifact to keep (not verifiable from code); the guard is the record, its owner and its upkeep.

Keep an AI risk register that identifies, analyses, evaluates, treats, and monitors each AI system's risks through its life cycle, including after deployment.

A written AI risk methodology (AI-specific risk sources such as data drift, bias, misuse, opacity, and third-party model changes; likelihood and impact criteria; risk acceptance thresholds) folded into the organization's existing risk process, plus a per-system register entry with owner, analysis, evaluation against tolerance, chosen treatment, residual risk, and the monitoring signal that would show the risk changing. The system owner keeps it, with the risk function, and updates it at design review, before release, and when monitoring or incidents show a change. Link register entries to the evals and production monitors that implement each treatment.

Where it goes: 12 repository artifacts, 13 tests and evals, 10 logs and telemetry.

Example (AI risk register (repo record)), before:

# risk-register.yaml
- system: loan-assistant
  risk: model might be biased

After:

# risk-register.yaml
- system: loan-assistant
  id: R-07
  source: training data under-represents applicants over 65
  analysis: {likelihood: medium, impact: high}
  evaluation: above tolerance
  treatment: reweighting + subgroup error-rate eval (evals/subgroups.py)
  residual: low
  owner: credit-ml-lead
  monitoring: dashboard approval-rate-by-age, alert at >5 point gap
  last_reviewed: 2026-09-15

Control: AI system managed without a documented AI-specific risk-management process. The same guard addresses 1 item. Engineering guidance, not legal advice.

Rule id iso-23894.ai-risk-management-process · review status: tier c citation only