[nybsop]

One issue that I've noticed with their system is that usually when a person posts an image, they take a random guess at the ID, and then all their friends will dog-pile on agreeing with it. There are so many misidentifications on there and not enough experts to correct it.

[sohkamyung]

I believe some of the 'random' guesses is actually the background AI system suggesting an identification.

Depending on the organism, it mostly works until it doesn't. :-) For example, in my region (Singapore), it's suggestions for mammals, birds and butterflies is mostly correct at Genus, Order or Family level.

I do agree that some identifications could be better and try to help by putting comments when I see something wrong (e.g., "this species is not native to the region"). But even putting in the wrong ID is better than not putting an ID and leaving a sighting as "Unknown". There is a huge pile of unknown identifications that nobody looks at because the experts usually filter by the organisms they know about.

Here is an old (2016) article about the AI recognition system being used by iNaturalist [1].

[1] https://www.inaturalist.org/pages/computer_vision_demo

[WilTimSon]

Could it be remedied by getting more experts to join? In cases like that, the hivemind is stronger than accreditations. It might just be a good idea to have some kind of moderation that can overrule people. Although that also adds the risk of alienating people who're very adamant about their identifications.