Is Ultra Hal True AI?

Realities & Misconceptions about Artificial Intelligence, putting it in perspective.

Article by: Don Gillett

What is considered to be real or true AI has been debated in attempts to define artificial intelligence for decades by the computer science community. The term "Artificial Intelligence" was just a phrase coined by computer scientist John McCarthy in the 1950's who defined it as "making a machine behave in ways that would be called intelligent as if a human were so behaving". This definition was pretty vague and left the door open for argument as to what can be considered true artificial intelligence, and to this day has yet to be settled or absolute in it's definition. For the time being, many have agreed that Artificial Intelligence, also known as machine intelligence, can be defined as intelligence exhibited or emulated by anything manufactured, usually hypothetically applied to general-purpose computers.

The phrase "artificial intelligence" has acquired somewhat of a bad name due to the large discrepancy between what has been achieved so far and the notions of what is considered to be actual artificial intelligence. The problem has been aggravated by science fiction writers and the movie industry which has raised the expectations of AI far beyond its current capabilities.

Presently Artificial Intelligence has been broken down into basically 2 categories "Weak AI" and "Strong AI". Strong artificial intelligence deals with the creation of some form of computer-based artificial intelligence that can truly reason and solve problems, a strong form of AI is said to be sentient, or self-aware. Human-like AI would be a computer program that has the ability to think and reason much like a human mind. Weak artificial intelligence on the other hand deals with the creation of some form of computer-based artificial intelligence that cannot truly reason and solve problems but in some ways still has the capability to act as if it were intelligent, but it would not possess true human like intelligence, such as self-awareness.

There are several fields of Weak AI, one of which is natural language which can be defined as specialized software or programming languages created for the purpose of distinguishing languages spoken by humans for general-purpose communication from constructs such as computer-programming that uses formal logic and mathematical logic to produce a determined or random intelligent response.

Ultimately Strong AI depends upon whether information processing machines can include all the properties and attributes of the human mind such as consciousness and reasoning. Weak AI however, is independent of many of the issues and attributes that constitute Strong AI and there can be no doubt that many of the features of modern computers such as multiplication or database searching could have been considered 'intelligent' not so long ago. To date, most artificial intelligence programming is computer simulations of intelligence based on predefined sets of rules, for example, an AI program will determine which database is the most appropriate to search for the most logical responses based on what information was put forth such as in a question, and in turn will also store information that is fed to it in a similar way so it can later reference that information when called for. Very little progress has been made in Strong AI, but depending on how one defines their goals, there has been good progress in Weak AI with programs like the Ultra Hal program, and we all must learn to crawl before we can walk and run and there's no doubt that Weak AI is the foundation or stepping stone to Strong AI. A case in point could be made with aero flight, not that long ago at the turn of the 20th century people were skeptical and even laughed at those who were jumping off cliffs with all kinds of contraptions in attempts to emulate flight in hopes that one day man could fly. Well a mere 100 years later privatized manned flights to outer space became reality all because of those crazy people who jumped off cliffs that are now considered the pioneers of flight.

Many of us have become victims of an impatient society that wants it all now, but has anyone thought of the implications that would arise if in fact an advanced form of Strong AI were developed tomorrow? Hypothetically, if we were to create a Strong AI with Ultra Hal that possessed all the traits of a human mind such as intelligence, reasoning, consciousness, emotions and self awareness, this would be considered by definitions as a sentience or sapience, a sentience is an entity with the ability to sense or feel, and a sapience is an entity with the ability to think about sensations, feelings and ideas, in usage sentience and sapience together imply some form or state of consciousness, which by definition would be perceived by many in the scientific community as being an actual living entity. If this were to happen tomorrow do you think we'd all be allowed to sit in front of our computers and play with our new human toy? I seriously doubt it, there would be implications regrading ethical and moral issues much like what is happening with cloning, and the Government would most likely seize control of such technology and it would tied up for years pending debates over the moral and ethical implications, and in all likelihood, also implementing regulations and guidelines regarding it's usage.

Although we fantasize about AI of this nature, I seriously doubt society as a whole is quite ready to embrace the reality of it, but fortunately we don't have to concern ourselves with that scenario for the time being because that type of AI is a ways off down the road yet and will be a gradual evolution. But in the mean time what we can do is make a choice to either be part of and support AI programs like Ultra Hal that will help bring us one step closer to creating a stronger AI, or we can choose to scoff at programs like Hal and alternately live out our fantasies and expectations of true AI at the movie theatres.

Is Ultra Hal a serious AI program or just an amusing toy?

The answer is both, and would depend on how Hal is being used and interacted with by the user. If a user is primarily feeding it relative and useful information that in turn will make Hal more knowledgeable and useful to the user, thus one might then consider it to be a more serious AI program. However, if a user is primarily conversing with Hal and feeding it more trivial information such as jokes or irrational nonsensical parameters for amusement, then in this instance Hal could certainly be considered a toy, in either instance Hal potentially will become somewhat of a prodigy of it's user. Ultra Hal comes out of the box not knowing very much with little personality, much like that of an infant child which needs to be fed information in order to learn and develop to the stage where it can make rational responses from what information it has already been fed by it's parents or mentors. For example, if an infant child had always been told that the sky was pink, it has no reason to believe otherwise and therefore would be the response one would get, as would be the case with Ultra Hal. An infant child has developed very little capacity for making rational decisions based on reasoning, the major difference between an AI program with machine intelligence such as Hal, and an infant human, is an infant human has the ability to react based on emotion or it's senses.

I believe many Ultra Hal users can eventually get to feel like they are actually conversing with another human. Does this constitute that Hal is real AI or just a means of tricking or fooling people into believing it is? If you take John McCarthy's definition ... "making a machine behave in ways that would be called intelligent as if a human were so behaving" is this not on some level what Hal is actually doing, behaving and/or emulating intelligence in way as if it were human? We can only hope Hal will continue to develop into a program that has an even much higher degree of human like intelligence that can eventually emulate things like emotions and senses such as sight, and of course we also want it to be able to do various tasks for us and just as importantly, have the ability to entertain us by making us smile and laugh. This would bring us closer to the ideology having a computer that is much more than just a machine and a mere tool, it could also become a form of entity capable of providing us with companionship or even friendship with a personality that can be molded or manipulated to suit our own.

In conclusion:

From the hypothesis and definitions put forth in this article which were based on research I've done and conversations I've had on the subject of "artificial intelligence" with people that have worked in the field of AI. Do I personally think Hal can be considered Real AI? The answer is YES! ... Do I think Hal can be considered advanced or what is referred to as "Strong AI"? No, and as far as I know, no such program exists at this time! .... Do I think Hal has potential to eventually become a "Strong AI" program? I'm an optimist and would never say "never", but that is ways off yet and it will depend on how much we believe it can and if we continue to contribute and support programs such as Ultra Hal.