I read two articles this week that, taken together, nicely illustrate what I think of as the state of the general public’s perception of AI.The first, “Chatbots Deliver the Worst Customer Service“, does a nice job of listing out specifically what is currently so frustrating about interfaces we generally call “chatbots”. I think it is summed up best in the article’s subheadline “Pretending to Be Human and Failing at It”. I think this sentiment applies to all non-human interfaces, everywhere. But I think it is also misdirected. We’re frustrated that we’re not getting the information we want, but we take it out on the “fakeness” of the experience.Compare that to another article I read this week, “Why the voice of the consumer is so important to AI“. Here we get some empirical evidence of why the voice interactions with Siri are considered so key to the overall experience. Again though, I’d argue that the underlying search mechanism is what is actually useful, but it is the pleasant voice and mannerisms that get the credit when you poll people.I think what often gets lost in general conversations about AI is the same sentiment that gets lost in every challenging application I’ve worked on: the functional performance is what drives people’s perceptions, but the aesthetic form is what they attribute their perceptions to.