Can AI Develop Awareness

This article is a continuation of a previous article entitled: Current Gaps in Artificial Intelligence.

In that article AI was defined as: “an electronic activity which emulates human brain thinking, only much faster and more efficiently.”

The article argued that AI only covers 50% of the way humans think.  For proper context, it’s recommended to first read Current Gaps in Artificial Intelligence. before proceeding.

This article discusses AI and “awareness,” or more specifically “self-awareness.”  The question being asked is: “When and how can AI develop awareness?”  Prior to answering this question, one must understand the meaning of “awareness.”

Science does not have a clear definition for the term. Putting both theological and philosophical conventions aside, a scientific definition runs on the spectrum between an amorphic to non-existent term.  This article strives to offer its own definition.  There is no way to prove the presented premise, as it is based on a subjective understanding of the term, but it does offer an adequate answer.

Turning our attention to a completely different topic, is discussion of a term called “Arrow of Time.”  This term borders on a philosophical discipline that describes “time” in terms of moving from past to future. The flow of the world, or human activity in the world, from past to future, is perceived by all of us intuitively.  However, as the Italian physicist Carlo Rovelli described: “The subjective feeling has no reference in physics.”

Laws of physics dealing with time, whether they are those defined by Sir Issac Newton or laws discovered by Albert Einstein such as the Theory of Relativity, do not address the Arrow of Time principle.  To the contrary, it’s possible to turn the theory on its head and use time from its endpoint to the beginning.  In essence, there is no connection between physics and our subjective feeling that time moves from past to future.  Or an alternative definition: “Time moves from cause to effect.”  In this case, where does the subjective feeling steam from?

Physics does describe something resembling the “Arrow of Time.”  It’s called Entropy. A scientific concept that is mainly associated with a state of disorder.  It is far from something intuitive, up to the point where it’s hard to believe that Entropy causes any subjective feeling.  For example:  For most people, heat is a source of readily available energy.  It is a very intuitive feeling, but from a physics perspective, entirely false.  Heat is a “graveyard” for energy.  Instinctively, when we think of energy, we think of work, or the potential for work.  Heat is the exact opposite. Energy turned into heat, will not turn to work.

At this point readers should be divided into two schools of thought.  Those who know and understand the term Entropy will agree with the statement.  However, those not familiar with the term are probably nodding their heads in complete disagreement. It is understood that for those not familiar with the term, this statement contradicts intuition.  For this reason, this example was presented.  It’s possible to prove the point, but it will distract from the ability to make the main argument of this article.

So where are we at this point? On the one hand as humans, we possess a very clear subjective feeling for the “Arrow of Time.” From the point where we describe time by its scientific definition, as described in Newton’s laws, physics is not able to explain our feelings.  That is because Newton can turn the concept of time on its head without any problem. On the other hand, the place where physics does explain the “Arrow of Time,” namely Entropy, contradicts intuition and our subjective feeling.

Such a paradox.

So where does our subjective feeling of “Arrow of Time” come from?  Of “time moving from the past to the future?” of “time moving from cause to effect?”

The truth is that there is no scientific answer for this. And by the same token, science does not have an adequate answer for “awareness.”

Readers should now begin to understand the point of this article.  There is no easy way to prove it, but the subjective feeling, of Arrow of Time, is awareness.  Just like our awareness is the subjective feeling we have regarding time.  And there is no real way to prove this theory, as science has not been able to define neither.  And if science is unable to do so, this article won’t be able to achieve it as well.

With that in mind it’s possible, using little effort, to find similarities between the two phenomena.  A point to be used as an example:  Self-awareness is contingent on our ability to discern between past, present and future.  The human capacity to think, the ability to plan our lives, including the awareness that life will end with our death.  Everything is interconnected with the understanding that time manages life.

There are countless other examples that readers can find on their own.  The easy way to find them would be the one that I used: trying to prove that this hypothesis is incorrect.  The more attempts made to prove it wrong, the more it will become easy to support it.  Readers are invited to accept this challenge.

The bottom line should be framing awareness using the following sentence: “Awareness is the ability to subjectivity discern the Arrow of Time.”

Now, that a practical definition of awareness was presented, let’s return to the initial question of this article: “Can AI develop awareness?”

The answer:  When software will be able to identify the Arrow of Time. 

For this to happen, uni-directional trends need to be understood within software.   In other terms, one needs to find or “create” non-reversable processes from which it will be able to draw direction:  First – Later; Cause – Effect.

It seems that a simple process is being described.  Or, is it?

Taking the following action as an example: “Permanently delete a file?” An irreversible process is being created.  However, it’s not truly irreversible as far as software is concerned.  If a file can be reconstructed, the software will be reinstated to its previous state.  The software – any software – is based on a method called "Turing Machine.” This virtual machine is reversible.  For comparison’s sake what is being described is equal to restoring an omelet to the state of a whole egg.  The latter is much more complicated than restoring a computer file.

Taking the RSA algorithm which is considered to be irreversible. It is still not really irreversible, but based on an action whose reverse direction takes much longer to achieve with the same computational powers.  And by bringing quantum computing into the picture, they make the processes even more reversable.

These are only two examples of many.  It is possible that there are non-reversable processes in computing.  Readers are challenged to find one.

More so, identifying non-reversable functions in software is only the first part.  The second and more challenging part, requires that the software detects and addresses the non-reversible action.  The bottom line is this: A software containing non-reversable processes, that in addition can identify and relate to them, will be able to possess a subjective understanding of the Arrow of Time.  This type of software will be able to develop awareness.

This will probably happen sometime in the future, but even with major advancements in the field of AI, this type of development is not expected to happen soon.

There is a school of thought that claims that AI will be able to replace the human brain only if it is able to connect to the physical world.  Not only by connecting sensors to software, but a physical connection, where software “pushes” something, able to sense how the physical world reacts to its influence (pushing back). This type of connection may bring AI closer to awareness.

In this instance the irreversible action will take place in the real world. The influence of the software on that world, and the world on it, will together, create an irreversible action.  Or in other words, Entropy can only exist in the real world and not in a software world. Hence, software must attach itself to the world to “experience” Entropy – an action needed to develop awareness.

In closing, awareness in terms of AI will be defined as the ability to subjectively sense the Arrow of Time, something software is far from achieving in today’s environment.  It will take time before AI will be able to develop awareness.  On the other hand, the subjective feeling that identifies the Arrow of Time is intuitive.  Meaning, it is primarily connected to system-one which is extremely developed in the AI world (as described in my first article).  However, it’s possible that full integration between both systems will be necessary. 

A final thought.  Many assumptions were presented herein without presenting much proof to substantiate the claims. Truth be told, the inability to prove was the main catalyst to publish.  I am working on some of the ideas presented in both articles, some in a more advanced state than others.  The main reason for writing both articles rest in my understanding that these topics cannot be addressed alone, so I invite readers to join my efforts to research and further develop these ideas.

Part of what I have presented here will be provable, and some may be contradicted. Most are in the “ballpark” and require more precise research. 

One thing is for certain.  Researching both will greatly contribute to humanity.