Sunday, February 24, 2013

Find Your Inner Vulcan

There is a famous Vulcan puzzle that has a purpose described by Voyager Lieutenant Tuvok as restoring the "structure of harmony". This game is used for the aim of going beyond Zen style Buddhist disembodiment, and instead allows Vulcans to focus on being rational and devoid of harmful emotions. The blocks have a variety of 3 dimensional geometries, that when stacked can easily tumble like a house of cards. Vulcans have the ability to go without sleep for up to 5 earth days, which it sometimes takes for them to complete the task. Other times it is done in a matter of minutes. The interesting thing about this puzzle is that there is no predefined form. It is assembled differently each time. So, despite the rather machine-like construction of these blocks, the outcome is creative. It is a humanoid’s attempt to be a machine, and his embrace of the outcome, that it is not merely a repetitive form, but something unique to that moment.
 As I look at these fictional characters trying to cope with the always present annoyance of emotion, I wonder if this quality the Vulcan is finding in this exercise is not merely mechanistic. What is it about creating unique forms based on that particular moment that makes a Vulcan easier for us as Trekkies to relate to? For some reason the one important human characteristic that I always think of as crucial in our ability to both improve ourselves and thrive as a species is empathy. The Wikipedia definition of empathy states: “Empathy is the capacity to recognize emotions that are being experienced by another sentient or fictional being”. At first glance this definition does not seem to apply  to a solitary act, but I would argue that in a very important way it does. The entire purpose of combining rationality with creativity is an applied discipline. When a meditation that results in variations depending on personal interactions of the day is completed, it is assumed that those interactions play an important role. Therefore to complete the task effectively the Vulcan must possess empathy. While this may be a stretch, it is allowed to be, as we are talking about characters in a TV space drama. The analogies that this suggests however, I think are transferable to us actual humans, not just pointy eared aliens.

This apparent contradiction between the logical mechanistic mode, and the creative empathetic one, is not only one to consider when assembly a starship crew, but also in a technology company, or even an artificial intelligence. I was attending a salon that the social scientist, and bestselling author Jonathan Haidt was a guest speaker for. The group was a libertarian organization called ReasonFoundation, whose name Haidt seemed to suggest, (based on his in-depth research), explains libertarians fairly well. It seems that empathy and rational decision making ability are inversely proportional. Libertarians seem to be the Vulcans of the political ideals spectrum (my words not his). This is in a way similar to very high functioning individuals with autism, such as many people with Asperger’s syndrome, who populate some of the best computer science departments and silicon value development labs. Haidt implies a mutually exclusive tendency between libertarian logical rigor, and a lack of empathy. While he certainly has a lot of research, there may very well be something more fundamental about the nature of humanity that he is missing.

Though I hate to jump around between fictional characters, real people, and robots I am going to do it anyway. Many people, who I promise you, are not all nuts, are considering the programming parameters for creating human-like artificial intelligence. I and many others think that computer technology is accelerating at such a rate that we will be faced with both practical and ethical questions about what and who these future highly intelligence machines should be. The former Singularity Institute, now known as MIRI, has been working on this very issue with full time researchers, and yearly conferences.

 Perhaps there is a something to the dichotomy of empathy and reason that we should consider when allowing our machines to become sentient. You notice I use allow, as I have decided to take a rather libertarian approach to artificial intelligence. As an analogy, humans have programming which is encoded in our DNA. A good generalized AI algorithm also has the equivalent, perhaps not written in ATCG base pairs on a biopolymer, but rather in binary logic, written in C++  and eventually printed onto silicon. We know how to create learning algorithms already. The IBM computer Watson, and most Google products, as well as thousands of others do it. These are machines that start as newborns with pre-programmed tools for learning  and as they age get smarter.  In some ways current AI does better than humans, and in some ways humans do better than computers. 

This is old stuff (well Watson was 2 years ago. I guess that is old in modern tech terms), and easy to understand. What we don’t know about human genomics, and equally don’t know how to do in computer science, is to find, or program that fundamental structure so that pure logical thinking, and empathy can co-exist. The reason that they should co-exist in AI may seem obvious. We would want machines that inherently have superior abilities, such as perfect memories, but also have the heart of an empath, or at the very least the heart of a libertarian Vulcan.

No comments: