Algorithmic Audit for Decision-Making or Decision Support Systems
Automated decision support systems (ADS) are machine-based systems that can make predictions, recommendations, or decisions influencing real or virtual environments for a given set of human-defined objectives. These systems are designed to operate with varying levels of autonomy. Over the last few years, ADS have grown exponentially in number and application areas. Currently, we interact with an increasing number of ADS, often without even noticing it. However, the lack of awareness about their use does not reduce the social risks if these systems are poorly designed or created without taking the necessary precautions. This guide seeks to serve as an instrument that helps raise awareness of the implications and consequences of implementing automated systems in making or supporting decisions that affect peoples lives. While audits have become a fundamental component in the growing field of algorithmic governance, they are not sufficient to mitigate the impact of a systems implementation and execution; they are basically a process to determine compliance with some standards. However, audits play a crucial role in impact assessment and information gathering and availability, for the entity itself and regulatory entities, and also for those potentially affected and the society as a whole.
| Main Author: | |
|---|---|
| Other Authors: | |
| Language: | English |
| Published: |
Inter-American Development Bank
|
| Subjects: | fAIr LAC, Governance, Auditing, Artificial Intelligence, Fiscal Transparency, Knowledge, Safeguard, National Accounts, Science and Technology, Gender, Justice Administration, O32 - Management of Technological Innovation and R&D, O33 - Technological Change: Choices and Consequences • Diffusion Processes, algorithmic audit;transparency;accountability, |
| Online Access: | http://dx.doi.org/10.18235/0004154 https://publications.iadb.org/en/algorithmic-audit-decision-making-or-decision-support-systems |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|