In the 1980s, they had to explicate their biases in order to make the algorithm's predictions correlate well with human judgment. Today we can just use neural n

In the 1980s, they had to explicate their biases in order to make the algorithm's predictions correlate well with human judgment. Today we can just use neural nets to learn our biases for us.