Project
Artificial intelligence - Quality requirements and processes - Risk identification and analysis for AI systems along the entire life cycle
Abstract
This document specifies an AI risk scheme covering the objectives
- reliability,
- avoidance of unwanted bias, and
- autonomy and control along the entire lifecycle of artificial intelligence (AI) systems that include Machine Learning (ML) components. It aims to serve as a basis for the identification of the AI risks which are present in a specific AI system and their analysis from a product perspective. Thus it aims to inform the parts of the risk management processes within an organization responsible for the identification and analysis of risks providing specific technical guidance for the mentioned objectives. The scheme maps risk sources to each objective and gives guidance on risk identification and risk analysis. This document does not specify the risk mitigation. This document is intended to be applicable to all developers, providers, and operators of AI systems
.
- avoidance of unwanted bias, and
Begin
2022-05-05
Planned document number
DIN/TS 92004
Project number
04301011
