#32 Can We Trust AI? About Fairness, Bias and Responsibility

In this episode, host Henrike Tönnes talks to Prof. Dr. Bilal
Zafar, Professor of Computer Science at Ruhr University Bochum and
an expert in trustworthy data science. Drawing on his experience in
academia and industry, Prof. Zafar provides valuable insights into
one of the most pressing questions of our time. Is AI fair?
Together, they explore the concept of unconscious bias in AI
systems, examining how it creeps into training data, design
decisions and large language models such as ChatGPT. What does bias
in AI really look like? Who is affected? And, most importantly: Can
it be fixed? The conversation also explores the role of synthetic
data and testing methodologies for bias detection, as well as the
shared responsibility of tech companies, researchers and
policymakers in developing trustworthy AI systems. Please note that
due to a technical issue during recording, the audio quality of
this episode is not optimal. We apologise for this and appreciate
your understanding — it's still well worth a listen!