22 April 2023

The Human Touch


The Human Touch


By Lawrence


Despite the search engines and what they dredge up from the internet, “human touch” does not mean being touched by a human.


If we want to refer to being physically touched by someone else we would say “someone touched me” or maybe, “I don’t like people touching me” or “I don’t like being touched by a doctor”. There, is therefore, a difference, from physical contact (touched by a human) and an idiomatic expression such as “human touch”. Indeed the most famous physical contact in literature is in the Gopel Luke 8:43-45-46, The Message, where Jesus asked, “Who touched me?” when a sick woman touched his robe.


The human touch idiom means, first and foremost, something like qualities that are unique to human beings which machines do not have. Machines might be made to mimic a “human touch” quality, such as empathy, warmth or understanding, but machines can hardly be programmed with an “if x then y” routine. Spontaneity and appropriateness to the situation determine what qualifies as a human touch and when to apply a human touch. (see eg: human touch at https://idioms.thefreedictionary.com/human+touch )


I introduced the idea in the context of our discussion on AI and relationships. Can an AI routine learn how to apply a human touch?


Another close idiom in digital technology is probably “user friendly”. The idea behind a human touch is to describe whether a machine interacts with people with the characteristics of a machine or with features humans can relate to. We all remember the early computers and mobile phones how they were set up as machines and were certainly not “user friendly”.


Machines that do not have that “human touch” or are not “user friendly”, behave, and come across to us, as if they were made by an engineering textbook. Up to an extent, this is understandable since early technology lacks that “evolutionary experience” to give us a human experience more like a human event rather than a “causal series of events” from a textbook.


We have all seen those cartoons of a caveman hitting his desired female with a club to attract her attention and then a modern man who bludgeons a desired female with two dozen red roses and a bottle of the most expensive Champagne. But this “evolutionary experience” also applies to machines; machines need to be efficient and robust in their environment because we want them to be user friendly.


Unfortunately, even today when people expect seamless applications and machines in both function and hardware (who reads manuals these days anyway?) there is a lot to be desired. One of the major shortcomings of machines that are in need of a human touch is the complexity of language employed to communicate with humans.


Not only are many messages not optimised for natural languages that people can understand, but are also long and full of technical terminology. And the worst case of absence of a human touch is the language used in error messages; Windows is one of the worst culprits here.


Another key difference about machines is that by definition they are linear systems following a cause-effect algorithm. Humans in general are more multitasking (in thinking at least) and are able to understand the overall picture without knowing all the details. Even more, we can exclude things in our thinking that we do not understand and can even recall and use information from experience independent of our present experience. Indeed, today we are more likely to interact with a machine using trial and error rather than read the manual first.


And even when we do read the manual we are more likely to read it in a trouble shooting frame of mind rather than to learn how to use the machine. I would go further and say that one of the differences between people in general from professional people (in their professional environment) is that professional people do learn about the complexities of a system and how to use that system. We prefer to shoot first and ask questions later; some even need to be shown what they have to do and say. They give the impression that they have no voluntary will to find out an answer to a problem.


One of the consequences of this gap between a human touch characteristic and a machine touch feature is that people do not venture into making the most of the features of a machine. Some, of course, do venture and explore the dark corridors of engineering and do successfully enjoy their machines and many of the in-built features.


So how important is this concept of a human touch? A complementary idiom to human touch is good bedside manners especially in a medical context. In the paper by Barry D. Silverman MD (Physician behavior and bedside manners: the influence of William Osler and The Johns Hopkins School of Medicine – DOI: 10.1080/08998280.2012.11928784) Silverman writes, “The accomplished doctor has a bedside manner that is humane and compassionate, empathetic and supportive.”


Of course, human touch is more a characteristic of acting like a friendly human being, whereas a good bedside manner in medical context must by definition include compassion and being supportive. I would argue that in human touch empathy is a sufficient condition although having extra characteristics would help. In his paper Silverman dates back the concept of bedside manner to 4th century Greeks.


Once again, we find the idea of professional people acting towards others beyond their duty and function very appealing. Although we do not hear this expression “good bedside manners” very often we do occasionally come across it in everyday life.  Unfortunately, we tend to use it in the negative to mean someone behaving unfriendly or even rudely. For example, the office manager is good but their bedside manners leave a lot to be desired.


Our idea of a human touch is when we go beyond the functional and duty. Human touch is that something extra only people who are empathic, caring and with feelings towards others who know how to bridge the gap between function and caring. So can an AI system go the extra mile when interacting with human beings?


Best Lawrence


telephone/WhatsApp: 606081813

Email: philomadrid@gmail.com



No comments: