Suppose you wanted to develop a serology test that could be used for this purpose:

to reliably identify a subset of people that have been exposed to covid19.

False positives identify people as exposed to covid19 when they haven't been, and would detract from your ability to identify someone as *reliably* exposed.

This is a nice thread that describes why the false positive rate has to be very low, lower than you might think, to avoid this problem. Especially when the incidence of covid19 is low. https://twitter.com/zbinney_NFLinj/status/1245789672833417217