Would it explicitly be taken into account? No. Would being the kind of person who'd get arrested at a protest be something our models could infer? Yes, though not at a high level of confidence. Would the specific fact of having been arrested have an effect? Probably not, unless they've got much better in the last few years. Would being the protesting sort affect lending decisions? If there was a correlation with non-repayment, yes, if not, no. We genuinely didn't care one way or another about that kind of thing, all we cared about was that the black box model could accurately predict how likely you were to default.