Ancestry.com or GEDcom are not "Sources" of original documentation

LegacyUser

David Newton said: Perfect 10 can be immediately distinguished because we are talking about Usenet in that case. Also note that said case did NOT touch on whether the uploaders of Perfect 10's content were covered by fair use at all. Note that Giganews removed copyrighted material from its servers when informed about it (the actual knowledge point). Therefore it also follows that Familysearch would remove copyrighted material from its servers if told about it.

Fair use is a multi-part test. You cannot just say "it's transformative" and stop there. All off the factors must be considered. So let's consider a typical scenario. User X goes to Ancestry and finds a record behind the paywall. They download the record image and then upload it as an FSFT memory. They do nothing to the image besides download it from Ancestry and upload it to FSFT.

So let's do a fair use analysis for that scenario. Is the use transformative? Absolutely not. It's taking an image from one online family tree site and reproducing it in toto at another online family tree site. What is the nature of the copied work? Factual and published. What is the amount and substantiality of the copying? Well if the Ancestry image contained multiple entries and the whole image was uploaded then more than absolutely necessary of the copyrighted work was used, but one image from a large sequence of images is not a large amount of the overall work. What about market usurpation? The use of the image on FSFT directly deprives Ancestry of income and is just about as usurping as you can get.

It's not transformative. Against fair use.
It's from factual, published material. For fair use.
It uses more than necessary of the work but does not use an enormous part of the work. Neutral.
It directly usurps the market for the original. Against fair use.

Two against. One neutral. One for. Oops fair use ain't lookin' so likely now is it? Importantly the two factors against fair use are ones that tend to be weighed quite heavily.

The only hope would be de minimis as a defence. If this were a particular person's own web site that would probably be viable. However FSFT is a multi-user website and thus the risk is not one user uploading a small number of images but of many users cumulatively uploading a substantial part of the copyrighted work. Hence my contention that the cumulative effect of lots of users doing this would have to be considered.

"Well, for one thing, that's not true. The process can be performed directly on the original without having to create an image in the interim."

Ah so you can put a book in front of a computer and it can read it just like that? Don't be absurd. You MUST have an image to perform OCR. As for the sophistry about human image perception in the brain just stop. You are resorting to a patently ridiculous strawman there.