Standards Development

Scope

This document defines a taxonomy of information elements to assist AI stakeholders with identifying and addressing the needs for transparency of AI systems. The document describes the semantics of the information elements and their relevance to the various objectives of different AI stakeholders.

This document uses a horizontal approach and is applicable to any kind of organization and application involving AI. V02/

Purpose

The goals of the new standard include:

- improving trust, accountability and communication among different stakeholders including partners in a supply chain, customers, users, society and regulators by establishing a consistent terminology around transparency of AI systems;

- guiding stakeholders about different elements of transparency with their relevance and possible limitations to different use cases and the target audience;

- serving as a basis for developing technology-specific, industry-specific or region-specific standards for transparency of AI systems This new standard can include informative annexes that would correspond to the objectives of certain stakeholders and tailored to their expertise. Each annex could include a “form” containing a subset of information elements from the taxonomy defined in the document. Examples of comparable proprietary forms, addressing different stakeholders’ needs, are Google’s “Model Cards” [[1810.03993] Model Cards for Model Reporting (arxiv.org)], IBM’s “FactSheets” [https://aifs360.mybluemix.net/] and Microsoft’s “Transparency Notes” [https://docs.microsoft.com/en-us/legal/cognitive-services/text-analytics/transparency-note-keyphraseextraction].