Humans have five senses: sight, sound, smell, taste, and touch. We already have cameras for sight, microphones for sound, and pressure sensors for touch. Can we consider AI a human if we develop smell and taste sensors to superhuman-level AI? Or, put the question the other way round, are humans just computers in the flesh?
A reductionist believes everything can be explained by reducing complex ideas or issues to their simplest parts. Psychology can be reduced to biology, biology can be reduced to chemistry, and chemistry can be reduced to physics. In this worldview, human behavior can eventually be reduced to a physical model, which can be programmed in computer language. In this context, some argue that humans are merely artificial intelligence in the flesh.
On the other hand, the holistic approach asserts that the total is more than the sum of all parts. An emergent property is a characteristic an entity gains when it becomes part of a bigger system. The bottom-up approach from the properties of lower-level components cannot explain this new property. We all know that a water molecule consists of two hydrogens and one oxygen atom. Nothing in the fundamental property of hydrogen and oxygen atoms describes water’s boiling and freezing points. They are the emergent property when a group of water molecules is bound together. There is also nothing in biology that can explain the social constructs such as “company spirit” and “group dynamics.”
Likewise, human consciousness is an emergent property of the human brain. No single property of a neuron is responsible for self-awareness, love, or hate. Nonetheless, the network of all neurons interaction in the nervous system generates complex human emotions like fear and joy, none of which can be attributed to a single neuron. Although the human brain is not yet understood enough to identify the mechanism by which emergence functions, most neurobiologists agree that complex interconnections among the parts give rise to qualities that belong only to the whole.
I do not believe in mind-body dualism. Dualism believes that the mind and the body are separated, and the mind is eternal and can exist outside the body. The mind is about mental processes, thoughts, and consciousness. The body is about the physical aspects of the brain neurons and how the brain is structured. This view describes very well the working of robots. The body is the computer hardware, while the mind is the computer software and data model. In some cultures, the mind is further separated into mind and spirit. The mind is responsible for logical thinking, while the spirit deals with spiritual things. Both views are not very convincing, and we should take the monist approach, which asserts that the mind and body cannot be separated.
There are two forms of monism; one rejects the physical body and claims that the physical world is a construct of the mind. An ancient Chinese philosopher illustrated this perspective with a dream where he became a butterfly and questioned whether he was a butterfly dreaming of being a human. This viewpoint likens the physical world to a computer simulation, similar to the movie “The Matrix,” where people exist under a computer-generated reality.
The other form of monism I subscribe to asserts that our mind and body are inseparable and constitute a united entity. Self-awareness and consciousness arise from neural activities in our brains. Our physical surroundings can impact our mental state, as we’ve all experienced when feeling tired, hungry, or affected by chemical imbalances. The physical body and the mind are strongly intertwined, as evident in medical cases where brain damage can result in personality changes. The emergence of human consciousness requires a physical body, which is why I believe in the physical resurrection of the body. We will delve deeper into this topic when discussing chance and necessity.
While it’s inevitable that some emergent properties will arise from artificial intelligence, the question remains: can human consciousness arise from AI? This question has no answer, but I am skeptical that AI can generate the same consciousness humans experience. AI could produce a consciousness similar to Spock’s from the iconic sci-fi TV show Star Trek, characterized by pure logic without emotion. Here are some reasons why I hold this view:
- The human brain is biological, while robots are electronic. They are two different subsystems, and it is unlikely that the same set of emergent properties can arise from two very different subsystems. Neurons in our brains are not like electronic switches. Neurotransmitters, such as dopamine, are molecules our nervous system uses to transmit messages between neurons or from neurons to muscles. Our diet and exercise contribute to how much dopamine we have. In a computer, there is nothing equivalent to these neurotransmitters. While computers can simulate those human behavior, computer simulation is not reality. The Incompleteness Theorem by Kurt Gödel stated that “for any formal system, there will always be statements about natural numbers that are true, but that are unprovable within the system.” In simple English, it means not everything is computable, and it is not because of a lack of computer power. There might be a lot of subtle effects happening in our brains that computers cannot simulate.
- Consciousness is not a built-in property of human beings. It is something that we develop as we grow up. Human learning and machine learning take very different paths. Every human is unique, and during the development of our consciousness, we become aware of our uniqueness. However, robots and AI are not that unique. If we load the same software and data model on two identical hardware, they are the same, so there is no or fragile concept of self for robots and AI. Therefore, machines might not have a strong sense of self-identity. Self-awareness is a critical aspect of human consciousness, and the consciousness that emerges from a system with a weak sense of self-awareness will be very different from human consciousness.
- Their different incentives influence the development of consciousness in humans and machines. Human life spans are limited to around 100 years, and our decisions are often shaped by our awareness of our finite time on earth. In contrast, machines can have an indefinite lifespan, with hardware that can be continuously upgraded and memory that can be transported from one place to another. While sexual identity is a critical aspect of human life, it is not typically discussed in church settings. Sexual desire is a biological urge that can lead individuals to act irrationally. While it is theoretically possible to program robots with sensors to simulate the pleasure of sex, the emergent properties that arise from physical and sexual desires may differ significantly from those that result from simulations. As a result, the consciousness that emerges from these two systems with vastly different incentives will also likely differ.
- Humans are often relationship-oriented, while computers are goal-oriented. Although humans can also have goal-oriented behavior, we can become easily distracted by unexpected events that arise during task performance. Unlike AI, the human mind has a vast network of associations, and when neurons are activated, it can bring up memories of other things that may cause us to divert from the primary task. While deep learning is also based on neuron connections, and related information may surface during task execution, the machine is programmed to identify and quickly discard any information irrelevant to the primary task. For example, if you ask Siri to call Jack and find a few Jacks in your contact list, it will ask you which one is the correct Jack to call. Once Siri starts the call, it will forget all the irrelevant Jacks. However, if you are a personal assistant, while going through the list with your boss, you might find a Jack you know well but last got in touch with a while ago. Once you have called the right Jack for your boss, you might call the other Jack during your free time.
I can go on and on and add more items to the list. Humans and machines have very different incentives and directives; even if AI can develop some consciousness, this consciousness will be very different from human consciousness. Do not expect them to be exactly like us.