What is the difference between computer memory and human memory?

Well, as you say, human memory is largely fallible. But it might not necessarily be as fallible as we generally think it is. By which I mean, there are ways of training your memory to be more reliable. Eidetic or photographic memories might be scientifically disputed, but there is the condition known as Hyperthymesia, whereby people can recall vivid details from any given day, even years in the past.

And, yes, human memory is by its very nature intertwined with, and often corrupted by, human emotion. But I’m not convinced that emotions aren’t just another form of data. Which is to say, I don’t think we can definitively rule out the possibility of a sufficiently complex AI developing the capacity for emotion.

I’ve encountered the idea (I don’t recall where — possibly in Star Trek; it seems like something that might pop up in Trek) that emotion is contingent on both organic biology and on a mortal existence. Why or even how, for example, would an immortal AI ever develop the capacity for fear? I think there may be something to that idea, but it also doesn’t rule out the possibility of AI emotions — it would just mean that we’d have to implant them into an organic interface.

I did read over your other article that you linked to, and I was going to post my comments about that here, but they’re growing rather long so I might as well post them to the other article instead.