No problem can be solved from the same level of consciousness that created it — Albert Einstein
This essay provides a framework for my readers to conduct small thought experiments of their own to come to their own opinion on whether a world ruled by robots is a possibility. My endeavor here is to not use any charts, data or expert opinions to guide my thoughts. This simplicity is intentional — I don’t want anything except my own personal biases (which I have many) to complicate the issue.
As of today, such a radically alternate reality doesn’t appear to be a possibility to me. These paragraphs outline why. I tend to believe there is a minuscule chance I am wrong.
Then again, I often laugh at my own self when I ask myself: who am I? The irony of the matter is that this joke of a question ‘who am I?’ is also an existential one. It holds the answer to what AI or even AGI can eventually become or not become.
A Few Underlying Assumptions
In the absence of another template, we are using the human brain as a template to model Artificial Intelligence and Artificial General Intelligence (AGI). That is my first assumption.
Second, I also assume that many people would already be familiar with viewpoints on both sides of this popular debate. There are very famous personalities arguing on both sides of the equation. Some claim machines can surpass human intelligence and even reach human level existence while many others think that singularity is not possible. I am not claiming to be anywhere close to these colossally gifted minds. Frankly, I am just a common man trying to make sense of this world.
Third, it is possible and will be possible to use algorithms to progressively convert data to intelligence and to automate many human jobs. There are three considerations here : the hardware or computing power (normal or quantum), the software or AI algorithms and the quantity of data (collected using sensors). Data and computing power will continue to grow while the algorithms (productivity) will continue to improve.
Having laid out these assumptions, below are a few simple thought experiments to guide your thoughts around this complex question. Hope they help you develop your own opinion on the matter.
It Is Not Possible To Go Beyond Intelligence And Replicate Human Creativity
Sophisticated algorithms using data can convert data to intelligence. The next question is can this intelligence neutralize the differences in creativity. Using stock trading or investment as an example, if everyone had equal access to data (which efficient market hypothesis presumes), can differences in human ingenuity still generate out sized returns (alpha) over the rest of the market. One can argue that hedge funds use more than publicly available data, more sophisticated algorithms and employ more intelligent people on average than the rest of the market to generate more returns than the market. In other words, they have more money at their disposal to fund talent to create sophisticated algorithms and to search for data that may not be available to the rest of the market. The question we need to ask ourselves is this: if everyone has access to the same data and algorithms through open source databases and algorithms, would someone else be able to generate a difference purely based on their creativity (strategy)? It’s a bit like Formula One but minus the differences in the construction of the actual car itself. In other words, if everyone had the same type of car, can the driver by himself/herself win a race? If the answer is yes, then Artificial Intelligence (AI) or even Artificial General Intelligence (AGI) may not be able to create unique symphonies, poems or pieces of art that a human could create based purely out of his/her own eccentricity or sheer luck/serendipity.
Second, if we apply this same logic to the job market, everything that can be converted from an art to a science i.e. anything that is either structured today or can be structured tomorrow with more datasets being added will eventually become a science. We need to ask ourselves: what jobs can be automated? Everyone agrees that the ones that are highly structured can be automated e.g. factory robots or robotic process automation. Most people would also agree that pure creativity which cannot be structured and is truly unique to an individual cannot be automated. The real gray zone is : what activities can be progressively automated if computing power, data sets and algorithms continue to improve i.e. what is present day art that can be converted to pure science. Will we ever see a day where portfolio construction will become a pure science? That is the million dollar question that nobody seems to have an answer to?
Finally, creativity lies beyond data and algorithms because algorithms are nowhere close to mimicking the behavior of millions of neurons firing simultaneously to produce that quirk of a piece of art that distinguishes Leonardo Da Vinci from Raphael. We cannot even seem to replicate the behavior of a single cell completely forget about all the cells and neurons in our body and the stroke of luck that is produced on many neurons firing simultaneously.
The Final Leap: It Is Not Possible To Convert Intelligence To Consciousness
The final impossible leap for AI is the transition to human level existence. What do I mean by that?
First, machines need to be made self-aware. As defined by Wikipedia - Artificial general intelligence (AGI) is the hypothetical intelligence of a machine that has the capacity to understand or learn any intellectual task that a human being can. What is still missing from this definition is self-awareness. Machines are not self aware yet. They don’t know they are serving us.
Second, to achieve human level existence, machines have to make a the ultimate leap from self-awareness to true consciousness. Consciousness is hard to replicate because humans don’t understand it yet. It is more complicated than self-awareness. Not only that, consciousness is too esoteric (spiritual even) compared to self-awareness. That is where the missing link is.
Lastly, the ultimate purpose for machines may not be the same as humans. Let’s conduct another thought experiment:
For most of us, we think more good deeds lead us to heaven and less of them compared to bad deeds would lead us to hell. For some other religions, more good deeds would lead us to a better re-birth. Ultimately, it doesn’t matter where you fall on the religious spectrum — all I am saying is that for many — their existence (atheists and agnostics not included for the sake of this experiment) is guided by a purpose even greater than creativity. All I want you to ask yourselves is this: do machines have a heaven/hell concept or for that matter the concept of machine Nirvana?
For atheists and agnostics, a negative view could prompt them to believe that machines will grow up to be controllers of humans. Perhaps, that is their ultimate purpose because on balance many of us will end up programming them to be evil. But, then again, even that purpose fails to take into account the entire diverse spectrum of human beliefs.
This brings me back to my ultimate irony / joke of a question: who am I? Until that question is solved — good luck to anyone positing a theory that says machines will be equal to human beings.