What’s Behind AI’s Bias and What Can Be Done About It?
Oct 17, 2025
This week, we're looking at the hidden bias in algorithms: how it forms, why it matters, and what it takes to build AI systems that are actually fair.
One Report
This report by the EU Agency for Fundamental Rights (FRA) explores how algorithmic systems used in predictive policing and hate speech detection can reinforce existing inequalities. It breaks down how bias forms, how feedback loops amplify it, and what’s needed to prevent AI from discriminating.
One Explainer
How does algorithmic bias actually work, and how can we stop it? This in-depth guide from IBM unpacks the technical roots of bias in AI systems. It also explores real-world examples across policing, healthcare, hiring, and finance.
One YouTube Video
In this TED Talk, Joy Buolamwini exposes the coded gaze, the bias in AI systems trained on non-representative data. She reveals how skewed datasets and untested algorithms shape who gets seen, hired, or trusted and what technologists can do to change it.
One Research Paper
Can improving fairness compromise privacy? This paper explores how tweaking a model to reduce bias, especially by changing how it draws decision boundaries, can make it more vulnerable to revealing who was in its training data.
One Article
Curious how today’s privacy-enhancing technologies are being used to tackle real-world AI challenges? This recap from Day 2 of Eyes-Off Data connects the dots between PETs, practical use cases, and what it takes to build responsible systems.
