If you're charismatic enough you can get away with saying how you really feel and you'll find that you make friends more quickly. I remember there was this time in the midst of a relationship ending and my father dying and after years of dealing with a painful medical condition I responded with "it's just another shitty day" with a rueful smile. Instant friends with the cafe worker. When things finally started to turn around for me and I said "things are great!" he was really happy for me.

It's a matter of reading the person and situation. Sometimes if the person asking the question is in a context where I can tell they don't want an honest answer I just say "oh, how are you?" I'd rather evade the question than to say "I'm fine" when things are shitty.

I agree with your view. Because the study result appears to rely on self-reports, it hinges on definitions of key terms by participants, especially what they consider a "lie". If the definition is like yours ("intent to deceive") that can be different than other definitions. Even then it's a spectrum because an innocuous 'white lie' is sometimes a social expectation, like: "I hope us being ten minutes late to the party wasn't a problem..." Not only is that sort of question often answered with a lie, it's expected to be. Depending on the context, the real-question-under-the-question is likely akin to "I hope our being late didn't bother you so much that it's going to create a problem significant enough to require addressing." A more expansive version of your definition might include "an untruth communicated with knowing intent to deceive another in a way likely to cause meaningful harm given the context." That definition would eliminate the example I gave above, however, it could also eliminate many of the 'inconsequential' lies the study reports pathological liars frequently make.

Reading the description of the study in OP article (but not the source paper), I felt there was no way to tell how reliable the result might be without knowing a great deal about how the study framed their questions.

Communication is about shared meaning. The shared meaning here is truthful; it's not so much about the words themselves. If your hearer grew up being taught to hear "How are you?" as a meaningless introduction, and you really wanted to know how they were, you would probably say "How are you really?". Isn't that redundant? Why not just say "How are you?"? Because you are aiming for shared meaning, not accuracy in absolute language.

This is an accurate portrayal of formal logic, but an inaccurate portrayal of language.