Artificial intelligence has resulted in a revolution of thought. Its immense innovation brings about self-reflection on our own imperfections as humans and the inversion, the transcendent perfection that a computer can bring from the uncharted territories of simulating the human brain.
Most, if not all, the decisions we make in life have the essence of bias attached to them. Think about the last minuscule decision you’ve made. Your lunch meal, song selection or the latest leased car. These decisions are colored by our own personal bias in the thought process. From a very early age, we have been influenced by actors that surround us with their own biases. Whenever we conduct the next decision, it is conjoined with our experiences that dictate our determinations.
As civilization evolved, the bedrock of developed societies demanded entities that would hold minimal inherent biases. The recognizable example, courts and the justice system. One of the earliest codes of law, the Babylonian Code of Hammurabi (c. 1754 BC), consisted of 282 laws ranging from contract litigation, inheritance and more. One of these codes mention the consequences of a decision:
(Code 5) If a judge has given a verdict, rendered a decision, granted a written judgment, and afterward has altered his judgment, that judge shall be prosecuted for altering the judgment he gave and shall pay twelvefold the penalty laid down in that judgment. Further, he shall be publicly expelled from his judgment-seat and shall not return nor take his seat with the judges at a trial.
It is a rational and necessary law code as constant changes in judgments would decimate the thoughts of a stable and efficient authoritative system. However, the consequences and pressures applied by such a code are immeasurable. To appease our own notion that humans can be unbiased, we appoint judges and panels of fellow juries to adjudicate other humans on matters of high importance. This is the best system we have come up with. Unfortunately there is one intrinsic human problem: we are biased, as we are, by design.
As such, those systems are flawed. Imperfection is what leads us astray. However, you might argue that this trait has allowed us to conquer new territories, venture into the unknown, survive under the most strenuous circumstances and even further advance civilizations’ developments. The survival and evolution of mankind is dependent on the trait and if you examine any conflict between humans, you will come to agree that it will always be a result of an aggregate storm of biases. Let’s go back in time to the dawn of the scientific revolution.
It’s the mid-1600s and a fortuitous event occurred. Under the shade of a genus Malus tree, an apple fell and an observant scientist presented a theory. These were the thoughts that sparked the theory of gravitation as described by Isaac Newton, resulting in a momentous event of enlightenment for humanity and science. A thesis came into shape about a force that everyone was aware of physically and at the least, subconsciously, but no one’s ever decided to describe it scientifically until Newton. You could argue that the discovery of gravity was on par with the discovery of the definition of the number zero, a concept early humans could not wrap their minds around but the integer revolutionized mathematics.
Bias needs the same enlightenment event to occur. In an age where artificial intelligence and data are already conducting decisions for our lives and livelihoods: loans/mortgages, job applications, autopilots, your next movie recommendations and many more facets of life, all controlled by software today. Computers are dictated by algorithms all the way down to the inner core, the chips that control the hardware. Some of these incredibly low-level algorithms already have weaknesses that in certain cases result in security vulnerabilities (ex: Spectre/Meltdown, targets caching algorithms in CPU chips) and plethora of other issues. Most of them, also, have the authors’ and companies’ biases embedded. As we move up the stack from the chip, to the different layers of software, that is where we see a familiar concept unravel; an aggregate of biases that only compounds and propagates with each layer. Does the company have a mandate to cut cost? The Boeing 737 Max software is a morbid example. Do you prefer to use the English unit system? Due to a failure to convert English to metric system, the $327.6 million Mars Climate Orbiter mission ended being incinerated in the atmosphere of Mars. Imperfections abound.
Computers epitomize perfection. They live by the laws of binary, and in the future, will prosper in the certain/uncertain quantum universe. As artificial intelligence reaches the levels of sci-fi utopian dreams, we will have entities that combine perfection with the dawn of unbiased thoughts. Can you imagine a future where your fellow human would be judged by an AI powered judge? Does that solve any kind of problem for society? Or is imperfection a quality that is required for human survival and redemption? We will find out.