I try to be patient. Really. There have been discussions about onboarding which would mitigate this. I will say, though, that I would rather someone use the CV – even if they are “blindly trusting it” – than upload with no ID at all. Some of these discussions where experts complain about data quality start to sound like “letting the perfect be the enemy of the good.”
If the tree has pods and bipinnate leaves, for example, the CV will suggest something in the Fabaceae. Even if it is the wrong genus, it will still get to the right family, and will be seen by an expert sooner than an Unknown.
Whether this will help with the problem of the CV mistaking plants for arthropods is a different question.