FAT Forensics: A Python Toolbox for Algorithmic Fairness, Accountability and Transparency

TL;DR

FAT Forensics is an open source Python toolbox designed to analyze and report on fairness, accountability, and transparency aspects of machine learning systems, aiming to mitigate potential harms and promote responsible AI development.

Contribution

The paper introduces FAT Forensics, a comprehensive Python toolbox that automates the analysis of fairness, accountability, and transparency in AI systems, filling a gap in existing tools.

Findings

Provides automated analysis of data, models, and predictions

Enables objective reporting for stakeholders

Open source and easy to integrate

Abstract

Today, artificial intelligence systems driven by machine learning algorithms can be in a position to take important, and sometimes legally binding, decisions about our everyday lives. In many cases, however, these systems and their actions are neither regulated nor certified. To help counter the potential harm that such algorithms can cause we developed an open source toolbox that can analyse selected fairness, accountability and transparency aspects of the machine learning process: data (and their features), models and predictions, allowing to automatically and objectively report them to relevant stakeholders. In this paper we describe the design, scope, usage and impact of this Python package, which is published under the 3-Clause BSD open source licence.