causal

AI Turbocharges Bias by Replicating Outcomes

An opaque AI that accurately predicts labor outcomes is a terrible idea because machine learning predicts by detecting patterns in actual outcomes; if those outcomes are generated by degree bias and occupational segregation, the AI will simply replicate and turbocharge the failures of the biased human system (garbage in, garbage out)—but AI applied to the right problem (e.g. teaching a worker the missing 15% of skills) could be extremely helpful.

causalpending

Speaker

Byron Auguste

Evidence Quote

“if the actual outcomes today are generated by all sorts of bias — and most obviously, for the purposes of this conversation, this degree bias — well, then the AI will say, "Oh, well, you're not going to probably be able to get that job because all our data says you won't."”

Source

Byron Auguste— Conversations with Tyler
Created: 6/12/2026, 4:53:35 PM

Your notes

Loading notes...