"...decrease the false negative rate, not increase the false positive rate."
The interaction between type I and type II errors virtually always have an inverse relationship. You cannot decrease one without increasing the other. It is, for example, one of the difficulties in making good medical diagnostic tests.
without dumping in a lot more money or w/e...
The final hiring decision is almost entirely entirely on the on-site interview, the phone interviews essentially act as a first-pass filter. And we don't do any sort of weighting of diversity in the final hire decision. From my experience, the phone screen has the highest false-positive rate. In this system the false negative is reduced considerably for diverse candidates, while the false positive remains mostly the same (since the on-site is pretty rigorous).
For what it's worth, while I can understand the author's opposition to this system I think it's an okay way of increasing diversity. I don't think this system has ever caused us to take on a diverse candidate of insufficient skill or experience. Does it mean that non-diverse candidates get filtered out on the phone interview step at a higher rate? Yes. But not so much that I'd consider it an issue - and this is coming from a (mostly) non-diverse guy. And while it does cost the company more money, in the sense that we spend more time interviewing some candidates than we normally would, it's money well spent.