Nobody is talking about actual LIES, just about the way certain data is represented. And no one has to accuse Amazon specifically of doing this, because this is ALWAYS done by EVERYONE. Which is why you almost never see meticulous, differentiated metrics, but only vague generalizations into "views" or "clicks" or whatever.
"View" can be anything, but someone who watches 1 episode, hates it, and never comes back is a VERY different data point from someone who watches all episodes start to finish. Saying "views are views" in response to that only reinforces what people are saying: that these kinds of simplified, one-dimensional metrics are extremely vague and extremely prone to creative interpretation.
As for account sharing... we don't know how many people do it, but we know it's more than 0 - and so the "views" are almost certainly higher than the number of actual accounts. By how much, we don't know; but lower, for certain. We also don't know if these "views" count repeats - so someone watching the show twice might show up in views twice, but would NOT actually generate revenue twice. And so on.
Absent details on the metrics we can't say for certain how many of their stated "views" actually translated into views tied at least potentially to revenue, but we CAN say with virtual certainty that it's LESS than the stated number.
The biggest distorting factor remains the conversion rate, i.e. how many people subbed because of the show and would not have otherwise (either new subs or continuing existing subs). That's data Amazon will never ever EVER make public, and it's difficult to speculate about. But it's VERY likely to be SIGNIFICANTLY less than the total number of views, probably by a factor of 3+ simply because that would already be an incredible conversion rate for any streaming service, let alone one where the sub isn't just about streaming.