How to use the Artificial Intelligence Act to investigate AI bias and discrimination: A guide for Equality Bodies
The Artificial Intelligence Act (AIA), the EU’s first legal framework for AI regulation, introduces new safeguards to address these risks. Equinet’s latest publication “How to Use the Artificial Intelligence Act to Investigate AI Bias and Discrimination: A Guide for Equality Bodies“ by Brent Mittelstadt, provides Equality Bodies with a practical guide to navigating the AI Act and its mechanisms to counter discrimination.
This publication has been developed as part of a project financed by the German Federal Anti-Discrimination Agency (FADA) to support Equality Bodies in investigating AI-driven discrimination. As artificial intelligence (AI) continues to influence key areas of life —such as employment, and public services— it is essential to ensure that AI systems do not reinforce biases or undermine fundamental rights.
Authored by Professor Brent Mittelstadt (Oxford Internet Institute), this guide is a key resource for Equality Bodies, equipping them with the tools needed to monitor, assess, and challenge AI-driven discrimination.
The guide explains how AI systems can introduce discrimination and the role of fairness metrics in assessing bias. It details how Equality Bodies can use new powers under the AIA, including access to technical documentation and testing rights. It provides practical steps to identify, document, and report algorithmic discrimination while emphasizing the importance of collaboration with Market Surveillance Authorities (MSAs) to hold AI providers accountable. Additionally, it outlines strategies to counter ethics-washing, ensure fairness measures align with legal standards, and address gaps in data and regulation.
This essential resource is designed for National Equality Bodies, policymakers, and legal practitioners seeking to ensure AI systems uphold non-discrimination and fundamental rights.
FADA Project 2023: Equinet’s Contribution to AI Technical Standards
In 2023 Equinet took part in a project funded by the German Federal Anti-discrimination Agency (FADA). Part of the project focused on contributing to the development of technical standards for the Artificial Intelligence.
Apart from the development of “How to Use the Artificial Intelligence Act to Investigate AI Bias and Discrimination: A Guide for Equality Bodies”, as part of this project, Equinet actively contributed to discussions on the development of harmonized technical standards for AI under the AIA.
Professor Brent Mittelstadt represented Equinet in Joint Technical Committee 21 (JTC-21), the body responsible for drafting AI-related technical standards under CEN and CENELEC. These standards will play a crucial role in shaping how AI providers measure and mitigate bias, ensuring that AI systems comply with fundamental rights protections.
Given the complexity and evolving nature of AI standardization, it is crucial that Equality Bodies play an active role in ensuring that AI bias standards align with non-discrimination law. “How to Use the Artificial Intelligence Act to Investigate AI Bias and Discrimination: A Guide for Equality Bodies” provides them with both practical investigative tools and insights into technical standardization efforts.