Not too long ago, on several occasions, Elon Musk had expressed his concerns about the development of Artificial Intelligence (AI). He even intimated that perhaps the world should have a moratorium on developing AI technology from the rapid rate of its trajectory. It is, according to him, "like summoning the demon" that could endanger humanity. It's not just him but other scientists and philosophers have issued similar warnings as well. The late Stephen Hawking, considered one of the brilliant scientific minds in modern history, cautioned, "The development of full artificial intelligence could spell the end of the human race." Too late now because the genie is already out of the bottle.
I don't know what Stephen Hawking meant by "full artificial intelligence" but let's dive into the stages of development of artificial intelligence as we know them today. Generally speaking, by consensus, there are supposed to be Ten Stages of AI development, theoretically. From what we can see today, AI is at stage 3 and likely at the early beginnings of stage 4. In my opinion, we've had the first two stages a very long time ago but it took a while to get to 3.
1. Rule or Knowledge-Based Systems
We - humans - can do 7X9 with very little thought because of the multiplication table that early on we've learned to memorize. Or, we just know to add 7 nine times in a row or vice-versa, 9 seven times. Rudimentary by today's standards, human programmers developed artificial intelligence in the basic calculators by providing certain rules and data by which to perform certain tasks quickly, sometimes repetitively, but unerringly so. A calculator when tasked to multiply 31,416 by 1,979, a tedious job for us to do, will still just add 31,416 1,979 times. But it does it at lightning speed. That is what is meant by a rule based system and providing specific rules, such as, multiplying two negative numbers give positive answers, but one positive multiplied by one negative gives a negative number, etc. It will know how to find the square root or cube root of any number at unbelievable speed that humans cannot match. At this stage AI merely followed the rules in its program.
2. Context-Based and Retention Systems
The thermostat in our home knows to turn on the AC unit because we set it to do so once the temperature reaches over the threshold limit of, say 72 F. It will turn it off once it is 72 F or cooler. The machine at the grocery checkout knows to honor a discount coupon only if you had checked out (purchased) the specific item. Television and the YouTube channel (owned by who else - Google), Amazon and other online services know a lot about what to show (or entice) you based on your viewing, browsing and (ultimately) buying habits. AI behind all of these knows what political party you lean towards favorably, or even guess unerringly the age group of the predominant viewers in a household based on their TV viewing habits. It will even know your ethnicity depending on just a few telltale signs.
The code writer gave it context rules and ability to retain everything in its memory.
3. Narrow Domain or Expert AI Systems
AI makes it possible for welding robots at the car assembly line to do its one and only job flawlessly with such sustained precision no human can duplicate; tirelessly, without taking a lunch break. That is all it will do and nothing else but no human labor can replicate even a fraction of its ability. The SpaceX booster rocket will lift off from its launch pad, soar up to 10-20 miles and return to the same spot precisely. That is a feat that requires adjusting for earth's rotation, wind and other factors. These classes of AI were designed to do specific tasks and excel at it. AI can play chess or Go or play Jeopardy to beat all human opponents but they are so narrowly focused to the exclusion of everything else. Alexa and Siri can mimic with their human voice but they are merely online information tools no different from a Google search.
4. Reasoning AI Systems
So many categories fall under this stage. ChatGPT will help a human write an essay, a speech or resume. A "true" reasoning system seems to be able to mimic thinking but doing it by, say, breaking down a problem in step-by step processes to arrive at the desired result. These are levels of algorithmic processes on steroids. To use a comical analogy, its train of thought avoids derailment because its reasoning can self-detect errors and has the ability to find new tracks if needed. It's not quite there yet to be classified as having general and super intelligence (Stages 6 & 7) but it is close.
5. Self-Aware Systems
This is a huge jump from Stage 4 but it seems a necessary prerequisite to levels 6 & 7 and beyond.
HAL, the computer featured in the Stanley Kubrick movie, "Space Odyssey 2001", was a self aware computer that took control of the spacecraft when a human astronaut decided against what it considered the "primary objective" of the mission.
The android in the movie, "Alien", was clearly aware of its place among all the human crew, and like HAL, it too simply had been programmed to adhere to the mission, unencumbered by ethical or emotional human responses to situations and conditions outside of predetermined parameters.
We will not go on discussing beyond Stage 5 at this point. Stages 6 to 10 are just too scary even at a level of merely imagining the consequences to our very existence.
6. Artificial General Intelligence
7. Artificial Super Intelligence
8. Transcendent AI
9. Cosmic AI
10. Godlike AI
However, I have expressed concerns about 6-10 stages early on already when exactly two years ago to the month, April 19, 2023, I penned, "A.I. The Sum of All Our Fears?" Below is a short quote from it.
"In 1970, I recall what I then thought was a B-movie science fiction, "Colossus: The Forbin Project". I am surprised to learn recently that it has garnered an 88% in Tomatometer ratings (Rotten Tomatoes); quite remarkable for a film made 53 years ago.
Movie Plot: "Colossus" was a super-advanced computer/military defensive system created in the U.S., located deep into a granite mountain somewhere in the Midwest; spearheaded by Dr. Forbin . It was powered by its own self-contained nuclear reactor, sealed from interference from the outside once the heavy doors were shut. It had the ability to repair itself and, more importantly, it was designed to totally deter any nuclear attack by other nations and respond autonomously. That is to say, it was capable of launching U.S. warheads on its own. Humans, including Dr. Forbin, can communicate with Colossus only via terminals linked by cable. The U.S. President declared that "Colossus was the perfect defense system".
Not long after it was operational, "Colossus" detected that there was another computer like itself, named "Guardian", located in the Soviet Union. I will stop short of crossing the spoiler alert line except to say that the two supercomputers started communicating with each other.
Elon Musk actually first made his warning on A.I. way back in 2016. He must know a lot more today so his concerns, along with others, couldn't be trivialized. Worth noting is that as early as 1940, Isaac Asimov, prolific and influential science fiction writer of that era but whose writings are still very popular today, must have had some concerns as well when it came to potential threats posed by "intelligent machines".
It was pure fiction fifty five years ago when that movie was made but Israel's iron dome - a defensive missile system - is a reality with proven effectiveness. The U.S. is planning to have a similar though perhaps far more sophisticated system given the size of the country to be defended. It won't be long before other military powers from Europe, Asia and the Middle East will create their own.
The new arms race will ultimately be about the level of autonomy these machines will be allowed to effectively respond as quickly as possible. The creation of the real "Colossuses" as in the movie is not far behind.
Meanwhile, along with these developments, we will be depending more and more on AI. From simple reference assistance to doing tasks for us mentally and physically, will potentially result in setting our minds and body in "bystander" modes long enough into states of decline. For example, our reliance on GPS navigation makes us lose our sense of direction without it. ChatGPT and constant reliance on the computer could make us lose or reduce reliance on individual creativity and critical thinking.
Perhaps, the greater threat is not from Colossus but from vulnerability of a weakened human mind. World military conflict will continue to occur because of our inability to learn the lessons from several thousand years of history. The next world war, if it were to occur, will devastate much of the first world. In a twist of irony, civilization will have to restart in places untouched by technology. Survivors will come from the interior of rain forests, remote islands and mountain tops where absence of technology will not have diminished humanity's reliance on creativity and physical prowess.
Theologians and philosophers tell us that Stage 10 (Godlike AI) is perhaps that point that God had forewarned about in Genesis 3:1-24
"Now the serpent was more crafty than any other beast of the field that the Lord God had made. He said to the woman, “Did God actually say, ‘You shall not eat of any tree in the garden’?” And the woman said to the serpent, “We may eat of the fruit of the trees in the garden, but God said, ‘You shall not eat of the fruit of the tree that is in the midst of the garden, neither shall you touch it, lest you die.’” But the serpent said to the woman, “You will not surely die. For God knows that when you eat of it your eyes will be opened, and you will be like God..."
In the end, we need to embrace and be grateful for the idea that of all living creatures we are the only ones given the ability of self awareness. With that we are also the only ones to be aware that there must be a Higher Power behind all of creation. Of course, the idea of free will, which is perhaps the one other ability we are endowed with, allows the reader to choose a different alternative thought process.