Americans often say 'Do you see x?' and 'I see x' etc. so it's not unidiomatic (in their dialects) nor incorrect. It's also appropriate if you're asking about a repeated event: "Do you see x [every morning]?" However, using 'can' is completely legitimate and versions with 'can' need to be added on all questions on this course involving verbs of perception. 'Can' usually doesn't need to be translated at all when it's used with verbs of perception in the present tense because in that context, it doesn't usually convey the meaning of being able to perceive something, only the fact of perceiving:
With verbs of ‘inert perception’ and ‘inert cognition’… there is little difference between being able to do something and actually doing it, so can tends to lose its distinctive modal meaning… With ‘verbs of inert perception’, furthermore, can not only loses its distinctive modal value, but has the additional special function of denoting a state rather than an event. As the Simple Present of these verbs has only an ‘instantaneous’ event meaning… the main difference between… I can see and I see is one of perception as a state versus perception as a (momentary) event
"Can you see x" can mean either "Vedi x?" or "Puoi vedere x?" and both of these can be translated into English as "Can you see x?"
It is not unnatural in Italian or English. I am a native speaker of English, and I use the "do you..." construction all the time, much more than "can you...". It is not a matter of always using "can" with perception; it is what is meant/what the situation calls for. This is my opinion.
Language is about meaning, not literal translation.
I am well aware of that. I tell people that all the time. However, if Duolingo is trying to teach you certain grammatical constructions, then for the purpose of learning the distinctions it's best to stick with strict (I never advocated literal) translations. The other thing I often tell people is that Duolingo is not meant to be a handy phrasebook. It's meant to teach grammar and vocabulary. And by glossing over the difference between "do you" and "can you", you are defeating the purpose of these lessons.
Duo is a computer program. And it's programmed to mostly ignore punctuation. So it really can't tell the difference between "See the fisherman" as a statement or as a question.
Besides which, Duo tends toward textbook-level language instruction, and while I agree that "See the fisherman?" is perfectly fine in ordinary speech, the fact is it's not "textbook" the way "Do you see the fisherman?" is.
In writing, a question mark turns a statement into a question. In speech, tone of voice turns a statement into a question. Verbs in English have do-support. Verbs in Italian do not. English needs a subject pronoun. Italian does not. We can tell from how the verb is conjugated -- in this case, vedi -- what the subject pronoun is -- in this case, tu, which in English is "you".
You see the fisherman.
Vedi il pescatore.
Do you see the fisherman?
Vedi il pescatore?