Behaving like robots

In my often frustrating dealings with ‘customer service’ help lines, I am encountering a new technology, the automated voice. The voice is almost always female and relentlessly cheerful. I understand the purposes of this innovation. First and foremost, of course, it is intended to save money, by enabling human beings to be replaced by automatons. Second, it is intended to allow a broader range of responses, for clients with no internet access or lacking the skills to use it, than systems requiring the customer to punch in a limited range of responses on a keypad.

In due course I believe such systems, if designed by human beings with creativity, humanity and intelligence, may represent a breakthrough. The artificial intelligence literature has long offered philosophical discussions of systems were human subjects were unable to distinguish, in conversation, between the responses of human and ‘artificial’ subjects. In his Foundation Trilogy, Isaac Asimov writes about robots that are more caring, intelligent and empathetic than the human beings they ‘serve.’

But we are a long way from that promised land. The present generation of robotic customer service mavens as a limited repertoire and is confused by anything approaching nuanced conversation in a normal tone of voice. The challenge for human beings on the other end of the line is to speak slow and distinctly, answering the robotic questions using a robotic vocabulary and in a robotic point of voice.

This is the paradox of robotic consumer service mavens at their present stage of development. The goal of their designers has been to create cost-saving systems in which robots behave like humans. But the results of their efforts, so far, are systems where, in order to get the information or services they need, we human beings are forced to behave like robots.