An ongoing theme in my articles has been how interaction with complex modern machines affects the psychological development of human beings. I have discussed how, for young humans, complex modern machines like video games, computers and smartphones become unconscious models for their behavior patterns. I have also explored how these complex modern machines mirror young human beings such that they, the young human beings, interpret their own behavior through the machine processes that they experience and with which they create an emotional connection. More recently, I have discussed how intense interactions with machines lead to a fusion of images of machine and human in the minds of people. And it seems that more and more scientists and engineers are trying to find new ways to blur the boundaries between human and machine on the path to making the fusion between human and machine complete. One of the means by which this fusion will be advanced further is emotion recognition technology.
Scientists are developing the technology by which machine software will be able to “read” human emotions on the basis of vocal expression and human gesture. Machines talking with humans on the telephone for sales calls will be able to adjust their presentation according to how this software perceives a person’s voice: the inflection, the pitch, the tone, the speed and the volume. For any human-machine interaction, the main purpose is to make the interaction as smooth, as frictionless as possible.
The purpose of this more frictionless communication is not just so that people and machines can actively commune. Rather, emotion recognition technology allows some machines to make ongoing adjustments in their presentation in order to control and manipulate the flow of the conversation. But there is no doubt that one result of these adjustments is that machines will appear somehow more human to humans.
With facial expression, other machines are able to read many different points on the face like the corners of the mouth or the curve of the eyebrows. Notice that this process is done by reading discrete points of the human face. Here we are again back to defined discrete stimuli. Humans are reduced to a series of discrete points, because modern complex machines operate on the basis of discrete stimuli. The continual flow of a facial expression, which is the basis of human intuitive interpretation, is completely lost in this process. There is a synthetic wholeness to a facial expression which is greater than the sum of its parts. This wholeness provides for a greater variety of nuance than can be simply reduced to discrete emotional categories like happy, angry, depressed. It is like the machine is breaking up the human face into pixilated fragments to facilitate interpretation.
The purpose of visual emotion recognition technology is for the machine to monitor the flow of engagement of a human with other people in a social situation, as, for example, a student in a classroom. The way the machine does this is by reducing the vast blendable continual emotional range in a person’s facial expression to a discrete finite series of emotional categories. The machine is programmed to respond according to which emotional categories are activated by its sensors. But the response is not going to be nearly as subtle, as nuance as that of a human being who responds intuitively to a whole flow of blendable continual emotional social communication. The machine is responding to categories or to combinations of categories – combinations of emotional pixels. The machine has a diminished infinity of responses with its limited awareness based on finite discrete categories of emotional signals from humans. The response of an emotion recognition machine will never be as fine-tuned as the natural response of one human to another.
Nevertheless, it will have a profound effect on humans, if they have to respond to these machines on an ongoing basis. The machines will stimulate unconsciously how humans start to configure their emotional responses to other humans.
In particular, they will start to act as a primary source of mirroring for people. People will start picking up unconsciously the gross limited patterns of reaction these machines display. In turn, they will start utilizing these patterns with other people, thus diminishing the quality of the rich vibrant interactions that are so important for making, preserving and receiving meaningful imprints. Human interactions will increasingly become implicitly more patterned and formulaic. More based on patterned figure responses in an emotional vacuum. People will not necessarily be conscious of this shift, but it will make relationships less satisfying and less strongly bonded.
And as organic connections start to suffer, people will become more remote from one another. And without the strong organic rejuvenation that comes from them, people will gradually become more machine-like. And as they become more machine-like as a result of all these intense interactions with machines – machines monitored by emotion recognition software – they will fall into a state of mind where they will be highly susceptible to being controlled and manipulated just like machines. Rather than having a coherent sense of self, the people who interact regularly with machines monitored by emotion recognition software will develop a fragmented pixilated sense of self.
More and more technology innovators are focusing on ways to blur the boundaries between machine and human. More precisely, the focus is on making machines behave more and more like humans. In the marketplace, we want machines to do more and more of the tasks of humans, because machines are thought to be easier to control and manipulate and they are cheaper to maintain. But as we work to bring machines up more and more to the level of human behavioral complexity, we start to increasingly bring humans down to the behavioral limitations of machines. As long as machines work primarily on the basis of discrete stimuli, discrete data, they will never be able to truly imitate the rich behavioral capacity of humans. For sure, they may develop a greater ability in those areas of behavior that involve solving rich complex puzzles like playing a chess game. But ordinary daily life cannot be reduced to a puzzle that has to be solved. Human interactions are too ambiguous and filled with complex ambiguous intentions. These complex ambiguous intentions are the basis of both comedy and tragedy. In other words, people don’t always follow a simple straight line from point a to point b in their relationships with other people. But emotion recognition software tries to reduce the complex vibrancy of human interaction to as much of a straight line as possible. It tries to make machine interaction with people smooth and frictionless in order to achieve a particular goal. Emotion recognition technology is simply another example of a technological innovation that will end up sucking the life out of life and converting people into robots.
The topic of this article was suggested by Dr. Jorge Cappon.
(c) 2013 Laurence Mesirow