causal

Models Drift and Acquire Bias Over Time

Machine learning models trained on a discrete dataset can pick up biases and behave badly as they are fed more data over time, and assessing this stability is one of the hardest things for an investor to determine.

causalpending

Speaker

Ash

Evidence Quote

“over time, as you feed it more and more data, will some biases start coming out in that model and will it start doing crazy stuff?”

Source

Investing in Artificial Intelligence— Invest Like the Best
Created: 11/19/2026, 2:19:59 AM

Your notes

Loading notes...