Skip to content
Artwork for Compiler
Compiler · May 8, 2025 · 34 min

Breaking Down AI Biases

Does your health insurance chatbot need to tell jokes? No. Does it need to be accurate? Absolutely. That's hard when biases get in the way. The introduction of bias into a model can be unintentional, but it can have significant consequences for those relying on its guidance. The Compiler team examines the ways bias can creep in, and what steps can be taken to address it.

0:00-34:27

transcript

No transcript — this publisher did not publish one.

show notes

Does your health insurance chatbot need to tell jokes? No. Does it need to be accurate? Absolutely. That's hard when biases get in the way.

The introduction of bias into a model can be unintentional, but it can have significant consequences for those relying on its guidance. The Compiler team examines the ways bias can creep in, and what steps can be taken to address it.

more episodes

All episodes