Another related thing I've been noticing: PageImages seems to strongly prefer images that are square-ish or portrait-sized. This makes a lot of sense for thumbnails in search, but not so much in Hovercards. Often suitable wide images, like logos, are skipped for this reason (even if they're PD-ineligible). But Hovercards can handle wide images just fine.

I've also seen a report where an image of the wrong person was displayed for a biography (there was no image of the person in question, but an image of another person was extracted from one of the navboxes.) Can't find the link right now.

Clearly, merely knowing the pageimage data is fine. A more realistic scenario would be that the 3rd party re-user does not realize that the returned page image may be non-free (Wikipedia is marketed as being free...) and then displays that image in such a way that it no longer passes the (legal) fair use test.

In T124225#1970207, @Deskana wrote:

In T124225#1970108, @Ruud_Koot wrote:

Could you give a concrete example of a situation where it would be appropriate to have PageImages return a non-free image?

I've already given one above; the lead image on the Android and iOS apps. That use is perfectly in-line with the non-free content criteria.

In T124225#1970059, @Deskana wrote:

However, there are many uses of this API that are in-line with that policy, or in fact many usages of it that are not on-wiki and therefore where the policy does not even apply.

In T124225#1965748, @Nemo_bis wrote:

This may well happen; which is yet another good reason to implement T95026: PageImages should be able to pick Wikidata's P18 as chosen image, so that editors disturbed by PageImages' choices can choose another (free) image without resorting to removing content from articles.

In T124225#1965527, @TheDJ wrote:

As I'm reading it here, 'product' basically takes the position of: "en.wp has some crazy rules and we are not going to accommodate them". I think that you'd be right to think that, but in my experience taking that position rarely ends well for said 'product'.

It will also further strengthen the "Ban all fair use content"-activists, so I'm not sure if "count" is a good measure, since too many deviations from the NFCC will just as likely lead to the full and total removal of all fair use content en masse from articles, in which case these articles also won't have lead images etc.

In T124225#1950749, @Jdlrobson wrote:

The downsides of such an approach would be images listed in https://en.m.wikipedia.org/wiki/Category:Copy_to_Wikimedia_Commons_(bot-assessed) which are free would be ignored.

Stop making this sound more complicated than it is!