In Design Theory class, we recently read 4 researcher’s writing on issues of problem solving.
Below are the axis labels to describe how I organized their positions on a 2×2 diagram in order to chart how each author views the importance of human nuance in problem solving. I explain each author’s stance in greater detail below.
My axis Y: Artificial Intelligence is capable of Ill Stuctured Problems, Artificial Intelligence is not capable of Ill Structured Problems.
X: Heuristics are good, Heuristics are bad.
Heuristics are experienced based techniques for problem solving, where the experience is so specific and personal that it often can’t be a good metric for accurate measure.
I wanted to see how these authors’ stances on these two issues relates to the importance of human nuance in problem solving.
Herb Simon has spent his career trying to figure out if computers can solve ill structured problems (ISP). He defines well structured problems as a criteria that need to be met in order to solve a problem- ISP are pretty much the opposite of that. The constraints around the problem needing to be solved are ill defined, thereby making the problem much harder to define and then solve. Simon was particularly interested in whether we could use computers to solve these ill defined problems and so spent a lot of research figuring out how people go about solving problems and also how computers go about solving problems. I thought that Simon would think that heuristics are good because he found that computers don’t have the capability of coming to illogical conclusions and heuristics are something that enable humans to do that. Computers can break down ISPs into some Well Structured Problems (WSPs), but in doing so, nuance is lost. So although Simon thinks that computers will be able to solve ISP at some point in time, at the time this article was written they couldn’t, and they can’t today.
Chris Pacione promotes the idea that design thinking should be a fundamental societal principle of thinking. Like mathematics. Everyone should know it and use it. It is should not just be in the tool box of designers. He describes how design research and insight synthesis was taught to nurses at a hospital and that they were able to design a solution to improve their efficiency and patient experience by 50%!!! So awesome. Because Pacione sees design thinking as what problem solving is (a human centered design process), he would not think that AI could solve ISP. He would view heuristics as indistinguishable in Contexual Inquiry and therefore not important.
Philip Johnson Laird talks about how for humans, figuring out a tactic or method for approaching a problem tends to make them rely on that method in the future and by nature then rules narrows the possibilities for use of other tactics towards problem solving. I charted him as being pro AI solving ISP, because computers don’t have that tendency and then therefore keep many tactics on hand to address problems. Heuristics only further narrow tactics, so he would be not be pro heuristics, either.
Amos Tversky and Daniel Kahneman are researchers particularly interested in heuristics. They talk about how stereotypes can give you an illusion of validity, or make you believe something must be true based on your experience. For example, if Steve is a mousy, meek, computer saavy, helpful person, we would stereotype him as a librarian instead of a farmer bc of what we stereotype librarians to be when in actual fact, he is much more likely to be a farmer. There’s way more farmers in the world than librarians. This in an incredibly human and flawed way of approaching a problem. Because computers remain neutral, I put TV&KM on the capable side of problem solving AI, but I also said that they think that heuristics are good- because they are human nature, and they allow us to distill information in an illogical way, which computers don’t have the capability to do.
How does all of this relate to human nuance? I think that if you look at my chart, you’ll see that the author’s stances are all really polar. The importance of human nuance in problem solving is the ability to make illogical connections. Computers can’t do that…yet. We can.
No Comments »