---
aliases:
  en:
  - AI safety case
  - AI safety cases
  - safety cases
concept_id: safety_case
preferred:
  en: safety case
relations:
  relations:
    broader:
    - ai_safety
    prerequisite:
    - risk
status: active
---

# safety case

A structured, evidence-based argument that an AI system is acceptably safe for a specified use in a specified operational context, typically making explicit the relevant hazards, risk controls, assumptions, evidence, and residual risk.

## broader

- [AI safety](../terminology/ai_safety.txt)

## prerequisite

- [risk](../terminology/risk.txt)

## Documents: mentions

- [アラインメント研究の自動化は、研究AIが意図的に妨害しなくても失敗し得る](../documents/bowkis_et_al.automated-alignment-can-fail-without-deliberate-sabotage.txt)
- [共通の重み・データ・訓練過程を持つ研究AIでは、研究結果の誤りや不確実性が相関し得る](../documents/bowkis_et_al.shared-training-can-correlate-automated-research-errors.txt)
- [control safety caseには、能力の十分な引き出しと、評価・外挿の保守性が必要になる](../documents/korbak_et_al.control-safety-cases-depend-on-elicitation-transfer-and-extrapolation.txt)

[Corpus index](../index.txt)