For everybody else, too, so far as I can see.

For Latinos in Arizona dealing with a new immigration law allowing official to stop anyone for no more reason than appearing to be an immigrant the potential accumulation of genetic information is staggering. Further, the accuracy of these results increases with the number of family members arrested again disproportionately affecting Black and Latino families.

Why does this disproportionately affect Black and Latino families? I think I'm missing one step in the logic here.

Finally, this effort would create perverse incentives. For every arrest, police will get more of this valuable information, which will incentivize arrests.

Do we currently experience people's being frequently arrested solely so their fingerprints can be taken? I don't think so, and it seems to me implausible that it will happen over DNA either. I think some basis for the conclusion here would be helpful.

Blacks and Latinos are likely to be the victims of these incentives.

You repeat that statement in this essay, and in fact it forms a crucial part of your argument, because it's why somehow DNA identification is "worse" than fingerprinting, but I'm not sure I understand it and I think it's based on a technical misunderstanding about what's in these "DNA databases." They don't contain representations of an individual's entire intact genome, that can be read as a book would be read. The information that is used to make these DNA-based IDs is much cruder. Imagine that you took a diamond and hit it with a hammer, then compared the fragments to the fragments you got when you smashed another diamond, to make a statistical judgment about how close to identical the two diamonds were, based on what was left after you smashed them. That's what's in the databases: a few summary numbers about some sequences that showed up often when you put some cellular DNA through the blender (it's not even nuclear DNA, given how much less of that there is, so you don't even have all the genetic material related to the individual, and much of what you do have is mitochondrial DNA, which is another genome altogether, and so forth). This sort of summary information might give you a good guess about whether two samples are from the same person (although there are also substantial uncertainties involved), but it doesn't tell you anywhere near as much about the person or persons as you seem to assume.

Now, at least in theory, for every entry in such an ID database there's a cotton swab somewhere, which might be stored in a way that isn't prejudicial to the long-term preservation of the actual material. It might at some time in the future be possible to take such a swab and get just a simple of the nuclear DNA, and actually amplify that so you could sequence it whole, and actually know what the whole genome of the human being was from which the sample came. But why would any police agency do that with all its samples, at greater or lesser expense, when it wouldn't do them any particular good? Today, under existing conditions, that's not technically feasible and if it were no one could afford it. They could do some specific genetic tests on each sample, but why would they, when each such interaction prejudices the sample's ultimate value as evidence?

So I don't know what the technical facts are on which you rely, or how that reliance affects your argument.

Further, family members with no contact with police may have their identities comprised as a result of the identification of family members as a result of DNA similarities. Finally, this practice will create perverse interests that will likely affect Black and Latinos in greater proportion.