causal

Early Design Standards Bake In Hard-to-Change Biases

Because technical standards get set very early and become hard to change later, failing to involve social scientists in robot design from the start—on issues like data security and the biases embedded when AI learns from biased datasets—risks locking in privacy vulnerabilities and perpetuated racial or gender biases that are difficult to undo afterward.

causalpending

Speaker

Kate Darling

Evidence Quote

“I used to think oh well the people developing the technology should just like create the stuff and then we'll figure out later on how to regulate it... but... standards get set really early on”

Source

66 - Living With Robots— Making Sense with Sam Harris
Created: 6/14/2026, 1:59:23 AM

Your notes

Loading notes...