Data Quality and Algorithmic Fairness

Every month, ETAI will be sharing a term or concept of the month that is related to ethical issues in biomedical research. The term of the month is part of our module's milestones relating to engagement and reflection on ethics with the AI-READI team.

Term of October 2023: Data Quality and Algorithmic Fairness

In today's digital age, algorithmic decision-making systems play a crucial role in fields like credit scoring and medical diagnoses. While they are often praised for being 'objective,' these systems can exhibit biases, mostly stemming from the data they rely on. Current methods for reducing bias in machine learning tend to focus on fixing the surface-level issues, rather than addressing the root causes. Taking inspiration from the idea that "garbage in, garbage out," it becomes clear that we need to shift towards a data-centric approach. This means improving data quality right from the beginning. Our research aims to use established data quality management practices from the database field to effectively tackle the underlying causes of algorithmic biases, resulting in more dependable and fair algorithmic systems.

https://arxiv.org/pdf/2212.10839.pdf https://dl.acm.org/doi/pdf/10.1145/3514221.3517841 https://dl.acm.org/doi/pdf/10.1145/3299869.3319901