What you've described sounds more like.... You've mentioned the issue about a human making a decision versus a black box doing it, or about having the courage to second-guess the answer the computer provides, but is that a privacy concern or is that a decision-making methodology concern?
I look back at the situation I'm more familiar with. The fact that a computer assigns a credit score is not where the privacy concern exists. It exists in this: did the person consent to disclosing the information? If they consented to disclosing the information, then decision-makers are free to either sort it out themselves or to plug it into a computer. That's not where the privacy issue comes about. It comes about through either a data leak, which is a separate issue....
Explain the privacy component to your concern around electronic decision-making.