The Long Road to Implausible Technology
Let me introduce real artificial intelligence to you, because modern culture has portrayed it in far too many ways. You may recognize the idea from a movie like Terminator, or Star Wars. In both scenarios, artificial intelligence (robots/cyborgs/droids in these cases) are used to wreak havoc in the world. This also may paint AI with a negative light in your mind. However, this should not be the case.
Artificial intelligence is defined by Britannica as, “the ability of a digital computer or computer-controlled robot to perform tasks commonly associated with intelligent beings.” These broad definitions don’t help with the common confusion associated with AI. First of all, let me establish that we all use AI. Every. Single. Day. Your phone’s calculator, Siri, Alexa, the chess computer, the computer in Call of Duty, you name it. Second, we all need to stop thinking of robots when someone mentions artificial intelligence. A robot is a house for an AI, not the actual machine itself.
In a famous essay, Vernor Vinge, a professor, and author, wrote about the concept of ‘technological singularity,’ the idea that at some point the technology that humans produce will become smarter than humans themselves. Current models predict that computing power will surpass a human in 2023, and by 2045… all human brain-power combined. Past this point, no one knows what may come next, there are simply predictions. Perhaps humans will become part computer, or perhaps we will begin to replace organs with artificial ones linked to a computer in our body.
The AI in video games is the lowest tiered AI, being named Narrow Artificial Intelligence, or ANI for short. ANI is a very goal-oriented system where the programmer can command it to perform singular tasks. The issue with this AI, and the reason we call it weak, is because it has many parameters and limitations. Have you ever picked up your phone, asked Siri a seemingly simple question, and she replies, “Hmm, I’m not sure”? I’m sure this is a shared frustration among Apple users, but it is simply because of Siri’s parameters. Although it may seem like Siri is another human at times, she is simply a program that is told what to do. Though ANI has been developing at a more advanced level in recent years because of Deep Learning. Deep Learning is why, when you search for a set of dumbbells on Google, a whole bunch of ads appear for… dumbbells. Google’s search engine collects your search data and then uses it to make advertisement recommendations for you; creepy. This is still years off of the next step up in AI: Artificial General-Intelligence.
This is where things turn from mundane to exciting. AGI is the theoretical next step in the development of AI, where it goes from weak to the equivalent of a human in terms of intelligence and processing capabilities. This AI would mimic human behaviors like being able to do mundane tasks such as putting a name to a face, or determining if someone is angry or happy based on how they speak. To achieve this, however, scientists will have to ramp up trials and expand their horizons on testing. One theory is that we will have to come up with a way where an ANI machine could rewrite its own code until it builds up the capabilities to match humans. Since the human brain is our model of building AGI, it is easy to see why we are having difficulty passing ANI. We simply do not understand the brain enough and how it functions to translate that to computer processing quite yet. Though, at the point when we do discover the secrets of the human brain, we will more than likely have achieved computational processing power that is more than capable of creating AGI.
Remember how I mentioned that AGI would most plausibly be created by creating an ANI that could rewrite its code? There’s another step past AGI, which is where things get a bit past human comprehension. Not only is the next step up a leap beyond our imagination, but it could also come merely days or hours after we first achieve AGI. This, of course, is Artificial Super-Intelligence. This is the point where it’s reasonable to think of Terminator; as in self-aware computers that exceed human ability and intelligence.
I’m sure you’ve already planted an image in your head about what an ASI would mean for the world; and I’m sure it’s most likely a negative view. This is because super-intelligence has been at the hands of science-fiction for far too long, depicting AI as the common enemies of humanity. While this is certainly possible, the same could be said for an ASI that helps humanity. The issue with an AI of this power would be that it would most likely understand things we, as humans, quite literally couldn’t comprehend. It would be like a Ph.D explaining quantum mechanics to a toddler (where in this case, humans are the toddler), but on an even more advanced scale. It would be hundreds, to thousands, or even millions of times more intelligent than even the most intelligent humans to walk the earth. Predictions hold the IQ of an ASI at between 12,000 and 15,000 (or even more). For comparison, an IQ of 135 puts a person in the 99th percentile of the world. Not only would these ASI machines be more intelligent than us, they would be better at everything we do, by a notable margin. Perhaps ASI would be able to answer one of the Theories of Everything, the Grand Theory of Physics which combines the General Theory of Relativity and Quantum Mechanics together to describe all phenomena in the universe. Or maybe ASI would be able to create methods of time-travel or interspace travel 10x faster than the speed of light. One of the most agreed upon propositions is that ASI, once achieved, will be able to make humans immortal. This, however, just touches the surface on AI.