Central Digital and Data Office and Centre for Data Ethics and Innovation
The Algorithmic Transparency Reporting Standard, jointly developed by the Central Digital and Data Office (CDDO) and the Centre for Data Ethics and Innovation (CDEI), serves as a national guideline for enhancing transparency in public sector organisations regarding their use of algorithmic tools and the rationale behind their implementation. The Algorithmic Transparency Recording Standard forms a part of the UK Government's National Data Strategy, that emphasises the need for greater transparency in algorithm-assisted decision-making within the public sector. The standard is structured into two tiers: Tier 1 provides a basic description of the algorithmic tool intended for the general public, while Tier 2 offers more comprehensive details on the data and workings of the algorithmic tool. The initial version of this standard was released in November 2021.
The Standard has been developed for UK public sector organisations that use algorithmic tools in decision-making processes affecting members of the public. Public sector organisations are encouraged to complete an algorithmic transparency report using the Standard for all algorithmic tools used by the organisation. Algorithmic tool has been defined as product, application. or device that solves or supports a specific problem by using complex algorithms. In particular, algorithmic tools that have significant influence on a decision-making process with direct or indirect public effect, or that directly interact with the general public are included in the scope of the Standard.
The UK Algorithmic Transparency Reporting Standard comprises two tiers. Tier 1 aims to provide a basic, concise description of an algorithmic tool's function and purpose in decision-making, catering to the general public's understanding.
In contrast, Tier 2 offers more detailed information for informed parties. It covers the owner and responsibility (2.1), describing accountability; description and rationale (2.2), providing granular details about the tool; decision-making process (2.3), explaining integration and oversight mechanisms; technical specification and data (2.4), delving into the algorithm's specifics and datasets; and risks, mitigations, and impact assessments (2.5), addressing potential issues and their resolutions. This tier is designed for civil society organizations, journalists, and public sector entities seeking deeper insights into the algorithms employed across the public sector.
