Conscious Free Choice and AI

The very essence of conscious free choice is first of all the freedom to choose an outcome based on feeling and not on coercion or even reason. Therefore, conscious free choice is a very good definition for conscious action because conscious free choice has a simple meaning as opposed to the many complex and contradictory meanings for consciousness. For example, there is a medical definition and all kinds of tests to affirm a conscious state having to do with the thalamus. At the root of all medical tests, though, is still the action of conscious free choice. In contrast, the term consciousness has broader psychological, philosophical, religious, mystical, and neuro-scientific meanings that seem to be beyond action. As a result, the concept consciousness is much less useful while the simplicity of conscious free choice makes it a very useful definition of conscious action.

Artificial intelligence has choice determined by algorithms and so AI choice is not free choice. Unless AI has a feeling driven by emotion, there is no possibility for conscious free choice. Conscious free choice is most apparent with the issue of free will, but there are many descriptions about the hard problem of consciousness that do not seem to involve free choice. Some people describe consciousness as self-awareness or introspection while others define consciousness with a set of questions and answers like the Turing test. Others say that how we feel about seeing the color red as qualia is consciousness and different from the fact of something that is red.

Philosophical zombies are hypothetical people without consciousness, but with otherwise normal behavior including conscious free choice. People argue that just the possibility of philosophical zombies shows consciousness as distinct from the machine of the brain. However, since people freely choose to think about philosophical zombies simply means that they are conscious. People also freely choose to think about the color red and so all of these descriptions of consciousness all are all equivalent to conscious free choice. Philosophical zombies look and act like conscious people, but have no consciousness. Since a philosophical zombie does have conscious free choice, such a zombie would then be as conscious as anyone else.

After all, people make free choices based on feeling and not really based on rational thought at all. Free choice outcomes are often then rationalized afterwards in spite of the lack of any rational precursor for free choice. Furthermore, the persuasion of another person can strongly affect a free choice. Persuasion can certainly affect how a person feels and therefore can affect a free choice, but that does not change the mysterious nature of free choice.

There are many people who claim that they somehow know the world is determinate and so that there are really no conscious free choices. They argue that the mind is a neural machine and that free choice is simply a result of the mind machine inputs and so not free at all. Even when free choice comes from feeling, determinism argues that the free choice was always a fated outcome and so that people simply have the illusion of free choice. Science very often makes the claim that free choice is an illusion of an otherwise determinate universe despite the fact that science does not have a measurement of free choice. Science admits that quantum superposition exists and limits the certainty of any measurement. Therefore, it seems especially perplexing that science makes any claims about free choice without a measurement of free choice.

A free choice outcome based on feeling does not have completely knowable precursors. In other words, we simply do not always know for certain why we feel that way that we do. In fact, we cannot predict free choice since free choice depends on feeling and we cannot always predict how we feel, either.

Artificial intelligence is often cited as a form of consciousness since AI makes choices in games like Chess or Go. However, AI is based on deep learning and so each choice of an outcome has completely knowable precursors. Moreover, neural emotion and feeling were not involved in AI choice, although one version of the deep learning process did involve competing policy and value networks. While policy played with more probability to win by as much as possible, value played with lower risk to win as simply as possible by one piece. Later versions simply played the computer against itself and so learned to choose moves from scratch over time to make better and better choices using both value and policy for both sides.

The AI computer is then all about choices with more or less risk, but not free choice since there is no neural emotion or feeling involved, just completely knowable machine logic precursors. Free choice also involves more or less risk, but with neural emotion and feeling. Feeling simply does not have completely knowable precursors, which makes free choice different from machine choice.

Now suppose the machine used quantum superposition as part of its decisions. With quantum superposition, the machine would now have free choices that would be as unpredictable as neural free choice and therefore also involve unknowable precursors. Quantum computers would then be the first true AI and should then be conscious with complex emotions and feelings. The AI emotions and feelings would then be as difficult to understand as are human or animal emotions and feelings.