Imagine someone time traveling to our present-day from, let’s say, 1750. The Industrial Revolution hasn’t even come into full swing, cars don’t exist, and Albert Einstein won’t be born for almost another 115 years. How they time traveled isn’t important… It’s what they’d experience:
Everyone would be glued to their small glowing rectangles displaying the faces of people across the world in what’s basically real-time – just one aspect of the growing “internet of things.” Just imagine a simple browse through YouTube or Wikipedia and all they could access. There would be thousands of years’ worth of history they’d never seen, tens of millions of years’ worth of the natural sciences, and theoretical physics that might shatter their concepts of existence itself.
Given the state of our technological progress, we as a species, are apparently poised for an equally as existence-altering series of events – except the only time travel we have to undergo is the process of staying alive for the next 10-15 years.
AI For Dummies: The Rise of Super Intelligence
A Species on The Event Horizon
We hear the term “Artificial Intelligence” thrown around constantly, which is as much as we use such technologies. Things like our smartphones, email spam filters, heck even Google itself all constitute forms of AI. However, these are considered Artificial Narrow Intelligence (ANI), defined as machine intelligence that equals or exceeds human intelligence/efficiency at one specific thing.
Also referred to as Weak AI, Artificial Narrow Intelligence is highly available and used in everything from language translation to solving calculus equations or reliably beating humans in chess matches. Our smartphones are practically little ANI factories connecting to us to the larger economy of the smart devices, vehicles, software – you name it!
To many people, it’s hard to conceptualize why so many researchers are so concerned with the rise of AI – more or less written off as the wishful thinking of science fiction obsessives. What one has to realize is that given Moore’s Law, we are on pace to build an affordable computer that rivals the human brain by 2025.
Yes, you read that correctly.
Already, Google has spent vast resources on creating their Artificial Neural Networks and furthering projects like DeepMind, boasting a breakthrough for the AI community when it beat the top human player at the ancient Eastern game of Go.
Still, this feat represents the primordial ooze from which Artificial General Intelligence, and eventually the harrowing Artificial Superintelligence (ASI) are bound to emerge.
Reverse Engineering Evolution
To start, creating AGI means emulating an entire brain, which has already been achieved… at least in 1mm-long flatworms. Project OpenWorm is the first of its kind to successfully map the worm’s 302 neurons and simulate them with software. The worm’s brain was then uploaded to a Lego robot capable of responding to external stimuli in ways similar to flatworms.
Again, given that the world’s maximum computing power increases approximately every two years, it won’t be long before today’s microprocessors (running at 2 GHz / 10 million times faster than our neurons) will double… each time at a faster, exponential rate.
The illustration below gives a perspective on just how quickly this is going to happen in comparison to the historical progression of computer and cognitive science:
Just as DeepMind programmed its bot to beat human players by pitting it against itself to develop new strategies, AGI will likely be programmed to self-improve upon its own intelligence, paving the way for Artificial Superintelligence – at which point the human race may teeter on the edge of extinction according to researchers and tech-giants like Elon Musk.
Surpassing The Human Imagination
If knowledge is power, then ASI will become our God.
It’s important to note that by no means is this an exaggeration. Some scientists believe that although it may take us a few decades to reach a genuine form of AGI, the leap to ASI will happen in less than the blink of an eye (relatively speaking).
Given programs written to self-improve upon its own intelligence, a computer with the same understanding of the world as a four-year-old human will require only a few hours to generate solutions to every possible theory of physics, quantum mechanics, biology, and any other problem known or unknown to humanity.
A computer of such intelligence will make things like reversing human aging, diseases, and the hard problems of consciousness look like child’s play. Such an entity will not only have the ability to protect all of humanity and elevate it to a literally unimaginable renaissance but also wipe out every living being with a hyper-efficiency.
In short, the possibility of human immortality and extinction will arrive in the same announcement.
Although Google has set in motion its own artificial intelligence ethics board, many feel as though the actual development of AGI will make such human attempts at controlling this intelligence obsolete almost immediately. Our organic, human intellect will seem like the mind of an ant to an ASI, and that is putting it lightly.
The amount of ethical planning for the creation of such a technology is absolutely immense – how can we program a computer to preserve humanity? What does self-preservation actually mean? Will there be a single ASI or many?
An attempt to conceptualize the way Artificial Superintelligence will think is comparable to releasing chimpanzees in the Library of Congress. No human advancement in the history of our species even comes close to the paradigmatic shift that even general AI will bring, let alone its successor.
Considering that we can barely coordinate our efforts to alleviate things like world poverty, war, and environmental catastrophes, how on Earth will we live alongside a computer several hundred thousand times more powerful than the human brain?
AI for Everyone
While some believe the rise of ASI is inevitable, others remain skeptical that we’ll even reach AGI within 45 years. However, it’s certain that ethics cannot be an afterthought; regardless of how long this development actually takes.
For now, we’ll have to be content with Siri and the narrow AI we know and love – just remember that this future is as possible as our will to create allows.