No, I think you're falling for exactly the problem addition of data creates: the discussion has shifted to the data, and jumped past the whole issue of immorali

No, I think you're falling for exactly the problem addition of data creates: the discussion has shifted to the data, and jumped past the whole issue of immorality. It's now "we don't steal because it doesn't work (under implied assumption of some eventual equilibrium if everyone steals)". BTW, under assumption of that "we" steal from "them", and "they" don't steal from "us", the data could support stealing. The data, conditions and objectives to maximize can also imply moral judgements, but are easy to launder in datasets, e.g. "better for the economy" sounds objective, but who is given access to benefit from that economy is a very moral judgement (touching hot topics like immigration). But the data is irrelevant, because the argument should stop at "we don't steal, period".