There is always the possibility that you will be incorrectly identified by an investigative technique, dna or otherwise. This is not a remote chance; it happens all the time. In my view, this does not increase that probability very much, but it massively improves the odds of finding criminals. Remember that the mere presence of DNA is not alone grounds for a conviction. There is still the matter of alibis, motives, and all those traditional avenues of forensic inquiry.

As to your questions, as far as I know, nobody has ever been murdered in a place where I've been. Maybe on the sidewalk somewhere, but that is not a scenario where DNA is going to be any help, unless the perpetrator leaves behind blood or skin on the victim or the ground. And I've never had sexual contact with anyone who was subsequently raped, nor anyone who didn't already know me for that matter. But it hardly matters whether those things have happened to me personally. What matters is how often that happens in the world, and what fraction of the time this leads to incorrect prosecution.

> this does not increase that probability very much

That depends[1] strongly on what question you are asking. Matching an unknown sample against a small number of suspects is a very different question than searching a database with millions (or billions) of records. The latter is almost guaranteed[2] to return false positives.

> it massively improves the odds of finding criminals

That's a strong claim that requires equally strong evidence.

[1] https://www.schneier.com/blog/archives/2008/09/dna_matching_... ( http://archive.li/M0SeV )

[2] https://en.wikipedia.org/wiki/Birthday_problem

>Remember that the mere presence of DNA is not alone grounds for a conviction.

That's awful wishful thinking. The general public is of the opinion that DNA evidence is 100% accurate and the government does a great job of continuing to present that farce. I wouldn't want to be the person who has to face a jury of their peers in the face of DNA evidence saying I committed a crime. Regardless of my alibi... I wouldn't trust the lay person to understand or get it right.

I would love to see some test cases of that. Call in an unknowing jury by normal procedure, conduct an apparently-real trial, present evidence that a) DNA means X did it and also b) X has three friends who swear to an alibi.

I'll bet X gets convicted.

Then that is what need to be fixed. Either make DNA information more accurate or better inform the general public the current state of DNA evidence.

People have been harassed and even arrested for sharing how bogus a polygraph is (by showing how to beat it). Good luck convincing people DNA evidence isn’t 100% accurate.

How do you propose to do that?

Otherwise you want to stop information ? You want to stop progress of technology ?