20190405.0430

One of the articles recommended for reading in my LinkedIN feed as I sat down to write this is Seth Borenstein's AP article "No AI in Humor: R2-D2 Walks into a Bar, Doesn't Get the Joke." The central idea is that humor is one of the most human of qualities, and that it depends on context--which artificial intelligence in its current state fails to comprehend. A number of technologists are quoted about the matter, noting the limitations on artificial intelligence despite its rapid advancement. Some warnings about the possible emergence of humor among artificial intelligences are also voiced. In all, the article does a decent job of presenting the present state of affairs and offering reassurance that at least some fields of human endeavor are likely immune to takeover by thinking machines--for now.
Of particular interest to me is a comment related from Columbia University's Allison Bishop: "computer learning looks for patterns, but comedy thrives on things hovering close to a pattern and veering off just a bit to be funny and edgy." To my mind, the comment echoes the words of another person from Columbia U; I am reminded of Asimov's comments about humor. There is no shortage of ideas about what makes something funny; that voiced by Asimov in his Treasury of Humor requires that a pattern be established and subverted. That is, there has to be some kind of pattern set up so that it can be deviated from--with the deviation offering the potential for humor. The simple deviation is not enough, of course; experience tells that it has to take certain forms and adopt particular content based on multiple prevailing contexts. How many jokes fall flat with one audience, only to have another rolling on the floor?
I have too much experience with such things to be comfortable recounting it. I want to be funny, to have people laughing because I want them to do so, not because they find me an object of scorn. But I think I try too hard at it; I certainly overextend in many of my attempts at humor, making connections that go farther than can be comfortable traversed or inserting a joke into a situation that does not admit of it. And it is because I do not take enough stock of context, focusing on the putative joke to the exclusion of its surroundings--not that the joke is necessarily good enough to merit that focus. So I wonder if the issue is not as much in the artificial intelligence as in other kinds altogether, kinds that are not necessarily accessible even to the "natural" intellect of a great many people.
Even if it is not, I suppose I have to face the fact that I am not good at being funny, even if I do often elicit laughter, and even if I can recognize being funny when I see it. I suppose it's like writing good stories in that; I can find them, and I can make sense of them, but I generally do not do well trying to make them happen. Clearly not, or I'd've sold more of them than I have yet done...