Big Data can Define Morality

The efficacy of a choice can easily be measured with great and copious data. But the morality of a choice is a different way of looking at decisions, and it's not obvious how morality be evaluated with a quantifier like massive amounts of data.

> If outcomes are good: moral. If outcomes are bad: immoral.

The words "good" and "bad" have the same problem as "moral" -- quantification is orthogonal to "goodness". Big data can only quantify, it can't judge something on moral grounds.

In 1600, everyone believed that astrology worked and was a personal asset. If we had had big data at that time, studies that only quantified would have revealed that nearly everyone believed that astrology was useful.

At that time, everyone thought astrology was factual, and everyone was wrong. No amount of data collection could have gotten around this structural problem.

Today, everyone seems to think antidepression drugs work, and if we judged the value of antidepression drugs using quantitative methods, we would have a clear winner. But as it turns out, antidepression drugs don't work, and a data-based, quantitative measure of their popularity is very misleading -- it only reveals the degree to which the placebo effect drives public beliefs.

It's the same with moral questions -- big data has a limited role if what's being measures is a mass delusion.