fAIr by design

fAIr by design is a research project involving eight partners from Austria (five companies, two universities and one NGO), focussing on the practical implementation of fairness requirements into AI systems. It is a three-year project, running from 2021 until 2024, funded by the National Foundation for Research, Technology and Development Austria (Laura Bassi 4.0).

Why is this project important?

The use of AI systems today has become pervasive, and impacts many aspects of our planet, society, and personal lives.  While AI-based technology has tremendous potential for positive improvement, applied blindly and without appropriate controls, it can be privacy invasive, detrimental to our environment, and might propagate inequalities and discrimination.

This has given rise to a multitude of ethical guidelines for AI systems - but so far, lacking practical input on how these can be implemented.  The goal of the fAIr by design research project is to bridge this gap.

What are the goals of the project?

The research project aims to develop, test and disseminate solutions that enable AI developers and user organisations to increase fairness and reduce the potential for discrimination e.g. against age, gender, race. A new fAIr by design process model and a toolbox will be developed, based on insights and lessons learned from bias mitigation strategies for fair AI in five use cases in the fields of healthcare, media, and HR; and using AI methodologies ranging from expert systems to neural networks.