[dehrmann]

I used to work for a startup that does ML for detecting payment fraud. Our untested theory was that the industry should be fairly recession-resistant because fraud increases in recessions, and fraud-prevention has measurable ROI, so you're probably not considering cutting back on it.

What'll be interesting is the meta-fraud of claiming false chargebacks or what the new fraud looks like because the shape of fraud is changing, and the model wouldn't have seen as much of it.