Can AI be Conscious One Day? Ft. @Phrenotopian

If you've heard anything about technology in the past few years, you would know that computers gaining human-like intelligence is a hot topic. But, many people such as Elon Musk and Stephen Hawking have cautioned against the misuse and danger of uncontrolled and fully conscious AI. Though, before we even think should we use conscious AI, we have to answer the question of whether we can even create a fully conscious AI. The concept of AI started with British mathematician Alan Turing. In a paper published in 1950, Turing outlined his logic on why a computer should be able to think like a human. Since humans make decisions based on logic and reason, computers could as well. The first of many breakthroughs for AI came in the summer of 1956 when Alan Newell, Cliff Shaw, and Herbert Simon presented their program Logic Theorist. This program was able to recognize patterns in the proofs in Principia Mathematica, recreate them, sometimes even making what the creators considered a more elegant proof. For the next 20 years, artificial intelligence found growth in a great amount of funding, more powerful computers, and improved algorithms for machine learning. While at Carnegie Mellon University, Geoffrey Hinton developed and applied the backpropagation algorithm, setting the stage for modern deep learning. The idea of consciousness does not come from the mainstream artificial intelligence that is incorporated in your home assistant or video game. These are examples of what is called artificial narrow intelligence, meaning a computer is able to perform a task at the level of a human, but a very specific task. Artificial General Intelligence, or AGI, is the concept that a computer can achieve human-level mastery in any subject or field which is chosen. Along with the ability to learn, the computer also has the ability to drop that subject and begin and succeed at learning something new. According to Aristotle, sentience and consciousness comes from perception, and perception of that perception and further extension of the perception. Aristotle called this higher-order thinking, and he believed that higher-order thinking was the mark of a truly intelligent being. According to Ray Kurzweil, the singularity is the point in time when the intelligence of computers surpasses the intelligence of humans. After the abilities of artificially intelligent machines grow exponentially and reach the singularity, the computers will hold more power than all the humans combined, and the intelligence which now holds control will start to seep into the surrounding universe. Kurzweil hypothesized that the advancement of artificial intelligence will continue exponentially, even after Moore's law becomes obsolete by greater understanding of the brain in a computational sense, improved algorithms for deep learning, as well as a gradual switch from an organic brain to portions of a digital brain. The singularity gets its name as the point at which man and machine become integrated. Lex Fridman, a professor at MIT, believes that AGI is an inevitable step in the development of technology. Fridman notes the biological inspiration in the perceptron, based on the neurons in the physical human brain. With the current architecture of artificial neural networks, the method of learning is very similar to the method of learning our brains use, in that it is a series of simple computations that occur at large scales. There is a sequence in which these computations occur in order to give us the ability to think, which we consider intelligence. As has been noted in many of the topics, one of the primary barriers to achieving human-like consciousness is the lack of in-depth knowledge of the brain. The human brain is known to have the densest cerebral cortices for all animals, providing the intelligence which allowed us to surpass other animals in survival. Unfortunately, as a result of this density, the human brain is rendered near impossible to map or model. By our current technology, this task is considered truly impossible. Furthermore, backpropagation, the algorithm on which all current deep learning algorithms are based, is nowhere near as efficient or comprehensive as the algorithms used in our brain whenever. For a more detailed analysis of why we may never achieve artificial general intelligence, check out the video by Frentopia. It takes a look at how AI may never achieve human levels of consciousness. But whenever I hear people talk about AI becoming self-aware and taking over the world to secure their own survival, I want to ask why? Why would an artificial intelligence want that? Why would an artificial intelligence want anything? In principle, it's just an extension of human wants and needs, and never its own intrinsic wants. But before you check out Frentopia's awesome channel, be sure to drop a like, hit that subscribe button, and turn on that notification bell if you wish to learn more about artificial intelligence and other fascinating aspects of our cosmos. And as always, stay tuned for more science videos.