069 - Artificial Intelligence (AI): Oxymoron or the Next Level of Consciousness?

Sundar Pichai, CEO of Google parent company Alphabet, has said  that artificial intelligence (AI) could have a more profound effect on humanity than fire and electricity.   Quite a statement.

New AI technologies are being produced, Including ChatGPT, that are conversational and can write better and communicate more clearly than most people.  And they provide fast, almost immediate, answers to any question.  While it is has not yet been perfected, and flaws have been noticed, the question has been raised as to whether such programs can self-learn on their own, and program themselves.  And importantly, when fully developed, whether they should be considered conscious entities.  Like a human being.   

There seems to be two camps here. One group believes that it AI is just an algorithm at the end of day, and does not possess anything beyond the information and formulas put into it by the programmers.  The other group feels we are on the verge of a creating a digital super intelligence, a digital god. 

What has philosophy and psychology to say about this?  This episode explores. 

Transcript

0:02

Hello, this is Gregory Nowak. This is the cunning of Geist episode 69. Welcome back. The purpose of this podcast is to explore philosophy, psychology and science with an emphasis on the great 19th century philosopher. George Wilhem Friedrich Hagle. In this episode, I'm going to be discussing once again, artificial intelligence. AI for short. And particularly in light of recent developments surrounding the new chat GPT software. This program has made the news lately for its conversational style and the speed with which it presents answers to questions. Many are claiming that this is a giant leap forward in terms of replicating human thought. And may even be the initial beginning of a sentience computer. One with actual consciousness. Now I did an episode almost a year ago on computers and artificial intelligence. That was episode 55. However, that episode was more of an introduction to the subject. And this was before all the talk about Chet. GPT in that episode, I spend time covering the computer revolution and the philosophical issues on whether a computer can truly be alive in the same sense as a human, I ended up being skeptical that a machine would ever be able to come alive, so to speak. But now we have seemingly developed a new tool. Which many claims is a giant leap forward in artificial intelligence. In fact, it has many people concerned that we might want to put the brakes on any further development of it. It appears to be able to reprogram itself. And some say, have a life of its own. If you can call that a life. And I know that's a very big if. Now there's an active debate going on in the media currently among developers, scientists, and philosophers. As to just what this represents or could represent in the future. And it seems to be two camps. One camp is that bottom line is just a bunch of algorithms and only does what it is programmed to do. The other camp believes that they have seen things in AI that go beyond a traditional computer. Some believe that the machine is actually showing signs of thinking on its own. So, this is what I'm going to be covering in the current episode. I must say that in preparing for this episode, it just seemed like I was getting bombarded with opinions on the subject almost nonstop right up until the time I started recording in the news and in social media. However let's back up. First, just what is Chad GPT? It is a software program produced by a company called open AI. In fact, you can go to the, you can go to open ai.com to read about the company and you can even try check GPT for yourself. I did try it out the website and I'll tell you what I found in a minute. The program interacts with a questionnaire in a conversational way. And according to the website, the dialog format makes it possible for Chad GPT to answer, follow up questions, admit mistakes, challenging, correct assumptions and reject inappropriate requests. Many scholars and reporters have been busy feeding it questions such as, how will you take over the world and things like that. And there's been somewhat of a media frenzy regarding it. Many are predicting the chat GBT as the potential to disrupt society as the internet did in the 1990s. Now there are other programs being developed beside chat GPT, Google parent alphabet has purchased a company called deep mind and is investing heavily in it. Tesla CEO, Elon Musk has stated that Google and DeepMind now have about three quarters of all the AI talent in the world, working for them. And what is striking? Is the reaction to cha GPT and these other efforts in all corners of the media and academia. Some claim it will replace up to 90% of all jobs. Others predict something far worse. It will actually take over the world is in the Terminator movies. For example, one fear article I found regarding Chet GPT claimed AI could cause a nuclear catastrophe. In a survey conducted by Stanford university. One third of all, AI researchers claimed an AI could cause a nuclear catastrophe by accident. And this was just through computer error, not a human interfacing with the artificial intelligence with perhaps. Uh, bad intentions. Elan Musk has also stated his concern that so much talent, is it Google? And that they appear not to be as concerned as much about safety as the good if humanity, but they're more concerned about profit than what this can do for, for humanity. He said in a tweet, Chet GPT is scary. Good. We are not far from dangerously strong AI in quote. And the television interview just last night, he said that we will be, we will need to be able to pull the plug. If something goes wrong. Musk also claimed that he had a discussion with Larry Page, the founder of Google's sometime ago. And Paige believed that it computer with consciousness must be considered to be alive. And never shut down, just like any other life. But we do, we may need to have an on-off switch and I'll be getting into this. This is obviously reminiscent of the famous scene in the movie 2001, a space Odyssey where Dave, the astronaut dismantles hell the computer was gone mad. Sundar Pichai CEO of Google parent alphabet said just this past Sunday on the 60 minutes television show. Then AI could be more profound and fire and electricity. Wow. That is some statement. What a time to be alive. And in the same interview, I said that we need to begin a conversation about what role AI should play in our lives. It is interesting than Musk is bringing up the term strong artificial intelligence and the quote I just read. There are two levels of artificial intelligence that are recognized in the field, weak and strong, strong AI or artificial general intelligence posits, a program that would be sentience. In other words, it would be self-conscious. It would recognize itself as its own being. More week I AI just performs thinking type tasks, both at the element of self-consciousness. A strong AI would be indistinguishable from the human mind, like a child. It may have to be taught things, but eventually it could learn on its own. Now at this point, strong AI is only theoretical. Some believe it is only a matter of time. However, before strong AI is achieved. And others believe it'll never be accomplished. I personally believe that we will be able to produce a machine that can achieve some 90% or so of the human capacity for intellect and reasoning. But we will never close the gap, never. And this is what I plan to explain in this episode, my reasoning for this. First, how would, you know, if a computer is sentience? Well, a common way to determine the intelligence of a computer model is the so-called Turing test. Named by Alan Turing in 1950, he was the inventor of one of the first functional computers during world war two. the test is essentially how well the computer could imitate a real person. For example, say there were two doors in front of you and behind one door was a live human being and behind the other door was a computer. You can ask questions of both and each would respond with a typed out response. The question is can the programming of the computer ever reached a point where it is indistinguishable from the human that you can't tell, which is the human in, which is the computer. If that point is ever reached, then some say true artificial intelligence would have been created. But again, if, if this is indistinguishable from a human, does that mean is alive itself? No, I do not believe it does. To me, this is just an example of weak AI doing logical tasks. And the touring test itself is not proof of a living Senti and computer. Some believe the Chad GPT is now past that Turing test. And it may just have done that. Let's. Examine a real world example here regarding halal. I logged on to the Chet GPT site and asked at the following question. What is the galleon dialectic? I will not read you the response, but first. Pretend you are an undergraduate philosophy professor and you were teaching a class on German idealism. And you asked the same question to your class and ask them to provide a brief essay on it. Say no more than 250 words. From what am I about to read? Could you tell. Whether it was from a real life student or the chat GPT program. Again, the question is what is the galleon dialectic? And here's the response quote. The galleon dialectic is a philosophical concept developed by German philosopher, George William fridger Cagle in the early 19th century. It is a method of reasoning that involves the development of ideas through a process of contradiction and reconciliation. According to Hagle ideas or concepts are not static, but are in a constant state of flux evolving and developing over time. This development occurs through the interaction of opposing forces or ideas, what she calls the thesis and antithesis. The conflict between these opposing forces leads to a synthesis, which is a new and higher level of understanding that incorporates the positive aspects of both. The thesis and antithesis. Hagle believed that this dialectical process was fundamental to human history and the progress of ideas. He saw history as a continuous process of development with each stage marked by a conflict between opposing forces in the eventual emergence of a new and higher level of understanding. The galleon dialectic has had a profound impact on philosophy, particularly in the areas of political theory and social philosophy. It has been used to analyze social and political movements as well as to explain historical events and cultural phenomenon. However, it has also been criticized for its complexity and for the potential for manipulation inherent in its emphasis on conflict and resolution and quote. Okay. Not bad. I would say that that answer represents a pretty fair description overall. And it just might pass is something coming from an undergraduate in philosophy who was studying Hagle. And this is what's getting everyone's attention. The chat GPT program may have achieved and passed some level of the Turing test. However, if this was a PhD program in philosophy, one might pick apart the response regarding the a galleon dialectic. First and foremost is the use of the term thesis, antithesis synthesis. As we often discuss in the Hagle study group on Facebook Hagle never use those terms and they really don't reflect what Hayden has in mind in his dialectical approach. So ablation is an important part of this, which is not addressed here. And it's unclear to me, just what is meant by the potential for manipulation inherent in its emphasis on conflict and resolution. For more on ablation and Hegel's dialectical process. Please see episode 30, six of the cunning of guys podcast. Yeah. Many others have tested chant GPT system and also funded wanting. The Washington post newspaper, for example, had 10 writers pose, nearly 50 questions to the program and they found about one in 10, about 10% were off-base or flatly inaccurate. They noted in the article that the biggest Brahm was the Chad GPT presents its answer in a definitive style, even when it's completely inaccurate. And many have pointed out that Chet GPT also produces phony references. For example. Not seem to live on Twitter, pointed out that he asked Chet GBT a question and it provided a reference as part of the answer. It's cited a journal the year, the author, et cetera. But when he put that reference back into Chad GPT, as a question, it said that reference does not exist. So obviously chat GPT needs some work. But many feel a new ability has been formed, which can be honed that will allow the software to calm the hive, mind of all of humanity. And be able to provide direct and meaningful answers to any type of question. And also to do much of the work that is now being done by humans. Some are using it on their jobs. Even now my son for example, is a middle school teacher and he told me he used it to help them develop a lesson plan in just a few minutes, which otherwise would have. Taken a much longer. I want to discuss next for a moment, how Chad GPT might do what it does. It obviously devours reams of data from all over the internet and books. And the programmers have figured out how to condense this information into a conversational style as we've seen. Now I have no idea how this actually works. However, I did spend my professional career in the field of market research and we used a. Statistical tool called factor analysis. It allowed one to take a large amount of data and find similarities of data points to other data points. This allows the computer to reduce the amount of data down to a smaller number of factors. For example, one could enter in thousands of bits of data to say how people responded to questionnaires both open and closed end questions. And then the statistical tool would boil it down to my specific number of factors. There could be five main factors or 20 or more depending on, on the researcher. And also there is statistical tools to help make that decision. If how many factors are best to use. So on scouring, the internet jet GPT may do something similar and reducing large amounts of data down to several factors. In the ha galleon dialectic, for example, a may have found thesis. Antithesis synthesis was used most often to describe dialectics and build its response around that. And that's why you may not be able to trust it entirely because it's just picking up what's out there. And with what's out there is wrong. Then Chad GPT is going to be wrong. And factors themselves can be used to segment the data. For example, there may be certain groups where some factors are most important and different groups where other factors are more important. Take the political arena, conservatives versus liberals. These are two clusters of, individuals and there's a tool called cluster analysis that identifies different groups. They respond differently. To the factors. So the factors are used to create different clusters and that allows you to understand the population even better. and it was interesting. This cluster analysis can even identify groups that have not yet been identified by the media politicians or anyone else, but they do exist. And on the less. The same can be done with the words themselves and sentences. Certain underlying patterns of sentence construction can be found. And this can be used to construct new sentences that have meaning. And here's where it gets interesting. It may be that humans also rely on these underlying patterns of speech to express themselves. If you really think about it. Much of what we think and say is repetitive. We use common sayings, common expressions, common emotions. Our brains may be similar and that we absorb much of what society has out there and we paired it back. And this. I believe is true to a large extent, but not completely. I believe there is also an intuitive, holistic way. Of thinking as well. What Hegel calls, reasoning. Or And German I've often referred to it here as right-brained reasoning, as opposed to the common either or thinking of the left brain. Our right brain can also perform pattern recognition and put concepts together and new ways to create new thoughts, new ideas. Chad GPT can only feed back to us. What is out there? Yes, it can learn how we most often put concepts together to form new ones. And mimic this process. But it cannot tell which combination will produce the Eureka moment. As we humans can do it may be able to predict it, but not replicate it. The true aha moment may be uniquely human. Essentially, I'm talking about an intuitive creativity here, the idea, the new idea, the better idea. But we, humans can go a bit further than just putting together different ideas. We can actually create ideas. This becomes very important in original music and I'm sure artificial intelligence will be able to compose music someday. In fact, a lot of music today sounds like it comes from a robot, but that aside I do not believe a computer will ever be able to compose one of several Bob Dylan songs, for example, or create something like Don McLean's big hit American pie. There's so much nuance in history and double meanings. In those words of those songs, that, to me, at least it would be impossible to produce these songs from a computer program. And the same goes for musical composition as well. Popular music uses repetitive chord patterns, but as I said is Dylan's words that were as genius. But Beethoven's ninth does not use repetitive chord patterns used thousands of times before. His ninth symphony was a unique, new and original piece of music and it stood the test of time. And there's also the performance aspect of music. The rolling stones produce great music. They wrote great iconic songs. And they have an incredible song book, but certainly part of their appeal is seeing Mick Jagger and Keith Richards on the stage. And with some artists such as mega star today, Taylor swift, how she looks and appears on stage, maybe more the driver of her success in the music. At most AI generated music would be a DJ at this Jackie. As human beings, we are in living, breathing bodies, and it is important for us to interact with other living, breathing bodies. That is part of the musical experience. Music operates in a deep, emotional and physical level that is grounding and binds us with others often, spiritually. That is why music is so important in religious services of all kinds. Interestingly, one of the original streaming music platforms, try to take the core elements of the music based on songs. The listener had previously selected and liked. And it produced a channel of music just for them based on an algorithm form for what they had previously liked through a kind of factor analysis. But it was a failure. Music is always the whole and not the parts it's not built up from the bottom from factor analysis. But it's an interplay from top down in the bottom up. Music itself is often heralded as spiritual something on a different dimension. Kurt Vonnegut said music is what convinced him that there is a God. Let me quote him cook. If I should ever die. God forbid, let this be my epithet. The only proof he needed for the existence of God was music and quote. Of course a computer could generate millions of different melodies, but it would take a human to pick out the great one. The computer would not know. You may have heard the story of. That given an infinite amount of time, a monkey typing on a keyboard, but eventually type out all of Shakespeare's works. But again, the monkey would not know how, or when this had been accomplished. Only the human being who knew Shakespeare's works would know this. Another interesting aspect of the AI revolution is the possibility of integrating personalities into the programs. We did the last episode on Young's personality types. It is interesting to think that at some point using young system programs could develop some eight different personalities using the system outlined in the last episode. Or 16 different personality types as in Myers-Briggs. So this would at least give the program a personality to some degree it could. Have a person choose one of several personalities to deal with. In the chat GPT. And make it more human, like. Some could be more emotional, some could be more rational and you'd end up with different characters, almost like the R two D two and C three PO. Robots and the star wars movies, or, or hell in 2001. But this discussion a personality, it brings me back to where I think AI can succeed and truly extend our capabilities and work can not. Marshall McLuhan said that all technologies are an extension of, of the human. All technologies are extensions of one of our senses or human processes, but Glen himself was incredibly prescient about things to come with the electronic revolution. And I urge you to listen to episode 21 of this podcast where I discuss McClellan. And the computer and Chet GPT are an extension of our thinking capabilities. And the computer revolution certainly helped our mental capabilities. Do children today even learn multiplication tables anymore with a calculator, always by their side. And we have spellcheck, we have programs to analyze our grammar and even sentence construction. And the internet has made libraries essentially obsolete. So so much is available just by the click of a mouse. And there's no doubt that AI and Chet GPT and its successors will dramatically enhance our logical capabilities as well. And it will certainly reduce the time it takes to perform many mental tasks. But rational, logical thinking is only one of several key functions of being alive human being. Many traditions have identified four essential categories of human existence. Using young system, they are physical, emotional thought and intuition. And we analyzed this in detail in the last episode. Number 68. And as we've discussed here before the Russian esoteric assists, George Gurdjieff called them the four waves. The, we have the faker physical, the way of the priest, emotional, the way of the Yogi rational and the way of the slide. One intuitive. We've also discussed how Gurdjieff use the horse drawn carriage as an analogy to the human being, the carriage, being the body of the horse, the emotions, the driver, the intellect, And the passenger, the master. If you will is our inner soul, our intuition, our purpose. And for more and Gurdjieff and this analogy, please see, episode 34. Regarding the physical function technology has greatly enhanced our physical capabilities. Obviously fire enabled us to keep warm and cook our food, the horse than the automobile than the plane. And the rocket ship a lot is to travel great distances to fly even to the moon. And unfortunately technology has also dramatically increased our ability to kill others, to kill our fellow human beings by the hundreds of thousands. Regarding the emotions we can connect with loved ones around the world through instant messaging and video FaceTiming, and we make friends. And if communities on social media, We need to get dates and meet people through online dating services. And then there's the logical thinking, which I've discussed. The computer and how it is increase the abilities of reminds dramatically and not just in computing power, but in information retrieval as well. But of the four functions. Intuition is an entirely different thing. Here we get into purpose. Into life itself. What drives us forward? We get into the collective unconscious. We get into spirituality. Now certainly as AI dips into the hive-mind of all humanity, it could even develop maps of our collective unconscious, which we ourselves are often unaware. Now I do believe that computers may be able to assist our inner selves. The master within, by understanding these undercurrents and the hive. Mind if humanity. By understanding these currents, we can gain new insight into ourselves and our motivations. It may help us decide what is. Not really unique, new or innovative in our goals. And w when we are in fact following just a script. This is the great benefit of AI to me. But here's the difference? The collective unconscious is a living thing inside Each of us. It is the spirit of memory that lives within us. All. Then we can assimilate this to a greater or lesser extent. As we live our lives. Now the computer could certainly simulate this process. this intuitive process by analyzing current events and comparing it to the. The hive, mind of all information. and, and coming up with. new new factors, but again, it would be acting mechanically through its own program and it may Ms. Key Contextual components. But even if it could keep up to date and integrate new data, how would it know what to do with this information? And here's a very key point. We ourselves, as a species. Do not have a clear idea of purpose that we all agree. To across the planet. We're not all on the same page as to what our purpose is here. Or if, even if there is a purpose. Spirit Geist is coming to know itself through our experience. Here is I believe in is Hagle taught. It is alive. It is the most alive part of us. And we fight and argue in history and folds slowly for the greater benefit of all. But the computer does not have the silent partner. It has to have a programmed in. But who is doing the programming? What is your objective? And this is where the rub comes in. Can we trust the leaders of Google to figure this out for us? I don't believe we can or should either. Does the Elon Musk. And just one more point on this ablation, which is the, so I just called it the driving wheel of a galleon dialectics. I believe sublimation relates directly to intuition. And please see episode 36 for more on Hegel's concept of simulation. So ablation is a living process in which an abstract notion as negation, it brought together. To form a new concept, lifted up to a higher level. And the key is that this process is not formulaic. It's not just averaging black and white to produce gray. It is a holistic process when the corresponds more to big picture thinking. Of the right brain. then the discursive either, or a symbolic logic of the left brain. Heiko called this The reason reasoning as being distinct from birth stand or the common understanding. As we discussed in the previous episode, this is what separates intuition from common judgment. And while Chad GPT might be able to model the intuitive part to some extent. I do not believe it will ever completely be able to duplicate this process. The reasons I have stated. Using the horse drawn carriage analogy. Of Gurdjieff's The human being operates as one entity, an integrated unity. Composed of four parts. The driver of the carriage may rely on ways or Google maps and other devices to steer the horses. And perhaps the driver can use computers to actually improve communication with the horses. But somebody must hold the rain. So living person. And the computer may simulate different objectives, different places to go, but the real direction and goal of the horse drawn carriage must come from the passenger, not from the driver with a supercomputer. Again, it is the passenger in the carriage, our inner being our spirit, our conscience, our life, our vitality that gives us the purpose. This can not be relegated to a computer. And this brings me to the last topic I'd like to discuss. That that of self-consciousness. And this is where a computer simulation cannot compete with the real life human being. In my view. As all Hagle students are aware, Hagle shows and is phenomenology of spirit that self-consciousness only comes about when a consciousness is confronted with another consciousness. There can be no conception of me itself without another me to compare. Now this sense of personal identity being a self can only come about when confronted with another. And I'm not sure how a computer could ever achieve this. I guess you could have two chat, GPT computers talk to each other, but again, they would only be blindly following their programming. They could fight each other for recognition, but again, this would be. A pre-programmed match. It wouldn't be a living, breathing matches in the master slave dialectic, the follows the self-consciousness section in the phenomenology of spirit. And the master slave dialectic, there was a fight to the death. And one gives up before either dies. The master is willing to die for recognition, but not the slave. How could this ever be programmed into a computer? There was real. Again, you could simulate this. It would be like fake wrestling on TV, but it would not be real. We are social beings. We've evolved from animals that are social beings. It is this living part that can not be replicated. Exactly. Now just some final thoughts. There was a notion of quality that we have discussed here before that term was coined by the American pragmatic philosopher, Charles Pierce. It refers to our actual experience of perception as in seeing the color red. The computer will never see the color red. It will be able to identify the vibratory frequency of the color red and distinguish it from other colors. But it will never actually see red. Like we do. The computer will never actually hear a musical note. Again, it can tell the vibratory frequency of that note being played and call it. That's a C or that's a. Or a flat, but it will never hear the note. Quality is what separates the human from the machine. And for more information and quality of please see episode 45. One final point. And the question I asked Changi. Petey about the galleon dialectic. It said, quote, it is also been criticized for the potential for manipulation inherent in its emphasis on conflict and resolution in quote Hmm. Manipulation. Another point. I Hagle study group member recently as jet GPT. What the reasons are to both study Hagle and not to study Hagle. One of the reasons given back to Nazi Hagle was as follows quote. Hey, those ideas have been associated with authoritative political regime, such as the Soviet union and Nazi Germany. While Hagle himself was not a supporter of these regimes. So may feel uncomfortable in studying his work in light of this association and quote. Now I have a major problem with this portrayal of hail. Both of them manipulation comment and the first code in the associates and with Hitler in the second response. These are the kinds of claims that take a life of their own. In the public consciousness and they stick. People are not going to take the time to study. Hey, let's see if this is true or not. I often get feedback when I posted about Hagle in different forums. some troubles. Oh, he, you know, that's. He that's where the Nazis got all their idea, some such nonsense. This stuff is very damaging and it shows. The problems with just how things are programmed and what comes out. Without supervision. AI could control the narrative and create the world that wants. it would be the programmers themselves, however, who are setting this agenda, not the computer. And this is where the danger lies. This is why we need some oversight and why we need an on, off switch. And we need to discuss this. as. As a human project. However, on the other side, by presenting us with what is essentially mechanical thinking. Finite mechanical thinking. AI can show us how much of our lives are purely mechanical, and this can enable us to be more creative in our own lives, in the lives of others. So there is promise here. So to summarize, I do believe that artificial intelligence is on the brink of a major revolution. That will certainly benefit the world. However, the way I see it as McCluen did, AI is just an extension of our own minds. I did not see a risk of it taking over the world or destroying humanity and some fear for the reason stated, however, people using AI may try to do just that. And for that, we have to be on guard. I also did not see artificial intelligence achieving true sentience or self-consciousness as a human being. And it's for the reasons discussed and may come close. And it may be indistinguishable from a human is in the Turing test. But again, it is just a simulation just to program. And it's best. AI will be a remarkable tool for the driver of the carriage, but it will always be the passenger in the carriage. Th that live passenger in the carriage that is calling the shots. And lastly, one further point. I realized this is a fast breaking development in, in the history of, uh, of humankind. And as more as found out. It's more becomes available. I myself will remain open to, to modify some or all of what I've said here. But this is where I stand now and I believe my essential points made here will stand. The test of time. So. That's a wrap for this episode. Thank you so much for listening. I will be posting references at the podcast Facebook page in a day or two, we had cunning of Geist. And please like, and follow that page. Cause I often post there in between episodes when we. We get very interesting comments there as well. And you can participate in the discussion of what we discuss here in the episodes. You can also follow me on Twitter at Cunningham Geist. And please tell your like-minded friends about the podcast helps spread the word. Thanks for joining me once again. And as always. This is Gregory Nowak. This is the cunning of Geist. See you next time.