These examples really aren’t difficult to come by if you look in the right places…
Just have a dig around in some of the taxa from the computer vision clean-up thread for commonly misidentified species and you’ll find stacks.
Here is a similar example I posted before :
https://forum.inaturalist.org/t/in-cases-where-the-cv-is-not-pretty-sure-of-anything-offer-a-suggestion-of-a-higher-taxa/14781/4?u=sbushes
Two new users both using autosuggest to take a blurry observation to species level. If users are new to biological recording… which many are… they likely won’t have any idea what constitutes a “difficult observation”. At the time, I took a look at the activity of one of the users as the OP has here… and saw they had just blitzed a whole bunch of observations in a similar way.
With no warnings in play, and the autosuggest itself offering species-level IDs for blurry indeterminate blobs…I think its fair to say the system as it stands actively enables this sort of situation.