Artificial intelligence is the field of computer science that focuses on developing computers capable of performing tasks normally performed by humans, especially tasks perceived to involve some human thought. Based on this definition, artificial intelligence imposes no restrictions on how the computers perform these tasks. A computer running software that solves a problem typically done by humans in a way that is completely different from how a human would typically approach the problem is still said to exhibit AI. In fact, there are 3 distinct camps on how to develop AI based on how much of the computing is modeled on human thought. Strong AI focuses on genuinely simulating human reasoning within machines in order to the perform the tasks. Weak AI focuses on just getting the machines to perform the task in any way possible. The 3rd camp is a sort of “in-between” for strong and weak AI, focusing on creating systems that are inspired by human intelligence. This 3rd camp is where most of the modern advancements in AI are taking place.
The big thing that separates AI from human intelligence is the fact that most AI are programmed to perform very specific tasks, while human intelligence is more general and capable of helping us handle the wide variety of challenges we face in everyday life. Take IBM’s Watson as an example. It’s designed recognize patterns and weigh the probability that it has properly recognized the pattern. While that may be helpful in winning Jeopardy or identifying health problems, pattern recognition is still only one task that true human intelligence performs on a daily basis.
This doesn’t discount the overall viability of AI. Platforms such as Watson, AlphaGo, and Deep Blue have made great strides in developing novel ways to push the boundaries of the types of problems computers can solve. While they are still limited by the fact that they were created for very specific tasks, there’s nothing stopping us from adapting them for more practical, commercial use such as Watson being used to aid in medical diagnosis. And while there haven’t really been as many great strides in strong AI, continued research in the field will not only help us create machines with the potential to rival human intelligence, but also ultimately help us understand how our minds work as well.
But what about those AI that have supposedly made the transition from simple calculation to what pretty much can be considered thought by passing the Turing test? While platforms like Eugene Goostman, the AI poet, and lamus, the AI composer, are all impressive in their own right, I do not think they should be viewed as heralds of the age of machines having minds. I do believe that the Chinese room is a good counter argument to the Turing test as its commonly employed. While Eugene Goostman is capable of fooling some judges with his generate conversations, it does not fully understand the meaning behind the words it produces (and can still be fooled if given the right input). While Apple’s Siri is capable to responding to human speech in a way the mimics a conversation, it’s still just code reacting to an input as opposed to a mind that genuinely understands the information it’s processing. I guess what I’m trying to get at is that the Turing test would be a much better measure of intelligence if it took affect and simulated emotion into account as opposed to just being able to carry on a simple conversation. If a machine is capable of somehow expressing joy upon seeing an image of the sunrise or feeling some form of companionship a la Her, who’s to say that the machine doesn’t have a mind of its own?
I do genuinely believe that some computing systems will eventually be considered minds, albeit minds that may function in ways different form our own. Think of it this way: suppose I built a human from scratch using the same biological materials, programmed its mind by implanting memories and a personality to where it would interact and react with the world like a natural born human, and then set it loose in the world. People interacting with my human would probably interact with it with no issue and would have no doubts that it has a mind with intelligence. Why should a computer be any different? What makes substituting the biological building materials for plastic and silicon and the mind with an OS so damning? The ethical implications of this idea would involve redefining what we commonly perceive as thoughts, minds, intelligence, and maybe even what constitutes a person. On a larger scale I think it involves us as a species stepping down from our existential pedestal to realize that we are not the “be all end all” of this vast empty universe.
codersongblog 