Ethical AI and Librarianship
A Resource Guide
Algorithmic Accountability for the Public Sector
|
Field
|
Description | |
|---|---|---|
| Title | Algorithmic Accountability for the Public Sector | |
| Type | Reports | |
| Creator | Ada Lovelace Institute. AI Now Institute. Open Government Partnership | |
| Link |
|
|
| Creation Date | 08/23/2021 | |
| Last Updated Date | -- | |
| Summary | Created by the Ada Lovelace Institute, AI Now Institute, and the Open Government Partnership, this report provides a survey of global algorithmic accountability policies in the government and public sector. In response to the growing use of algorithms to support public service decision-making, which can lack transparency and cause harm, regulatory policies have been developed to ensure algorithmic accountability. The report reviews a range of algorithmic accountability policies across countries, primarily from 2019 onwards, encompassing legally binding regulations, high-level principles, and voluntary guidelines. Examples of types of policies include: 1) high-level ethical guidelines, such as the UK Data Ethics Framework, which offer normative standards for public agencies; 2) prohibitions on certain algorithmic systems, such as ban of facial recognition technologies in USA local governments including Portland, Oakland and San Francisco; 3) public transparency mechanisms, including registries of algorithmic systems, source-code disclosure requirements, and explanations of algorithmic logic; 4) algorithmic impact assessments (AIAs) designed to evaluate potential harms before deployment; The report outlines six key factors influencing effective implementation of these policies, such as clear definitions of “algorithms”, appropriate policy scope, audience-tailored transparency. The report offers recommendations for policymakers to enable policy monitoring and evaluation, and identifies areas for further research in the field. The publication is available in two formats, an executive summary that presents key findings and a full report for more details on the findings. | |
| Topic | Ethical AI. | |
| Source and Link | Open Government Partnership. https://www.opengovpartnership.org/ | |
| Access | Open | |
| Accessibility | -- | |
| Audience | General. Scholars and Students. | |
| Platform or Format | Document (.pdf) | |
| Length | -- | |
| Geography | USA/GBR/CAN/CHE/MAR/Europe | |
| Language | ENG | |
| Description Date | 06/19/2025 |