the first think1 I usually do is try to reconstruct the original text, assuming it was translated from EN.
the next thing I would do would be to select an EN option if it was available because it would likely be easier to read (even if that is also a machine tl from another language; EN being my first language and all)
if it's in the context of personal communication, like a social media post or reply, and it has a literally translated idiom or a similar quirk, I might spend a few seconds wondering whether the intended reader will be able to understand it okay.
if it's in something commercial, I usually find it a bit sad. the person putting the text there might not have had enough resources to get it done properly, or might have a naive idea about the performance of machine translation, or might simply not care, and there is usually little hope that a better translation will appear
speaking of which, some of the microsoft online JA help pages have some utterly baffling machine-translated parts (like... the titles for starters. and it's baffling because a standard gtranslate job wouldn't produce anything so syntactically abhorrent) - you'd think that a company as big as microsoft would at least get someone to check the output, but it's as if they don't expect anyone to read it or are too arrogant to care.
this is more about machine translated text in general as opposed to something machine translated into JA specifically though. I think a native JA speaker or someone studying sociolinguistics/orientalism/japonisme would be able to talk about that better than me
-
gonna leave this typo as it is


