Thank you for your question.
The false positives I was talking about related to familial searches. This is a practice that happens in the U.K., where searches will be run against a convicted offender index from a crime scene and they look for close matches. That's a less precise form of search, so that is more likely to turn up false positives, or conversely, false negatives. I was talking about that context specifically, not matching of samples.
The technology has much improved since it was first initiated, so it is scientifically much closer when you're talking about a precise match. So I was talking about familial searches, which we recommend against.