Artificial intelligence is easily one of the most prevalent themes in all of science fiction.
The idea that a machine could exhibit the same level of intelligence and sentience as
a human being has captivated writers and audiences alike for decades.
From an ominous computer system in 2001: A Space Odyssey...
DAVE: Open the pod bay doors, HAL.
HAL 9000: I'm sorry, Dave. I'm afraid I can't do that.
...to superhuman androids in Westworld...
DR. ROBERT FORD: Turn the other cheek.
...this captivating sub-genre of science fiction has experienced a diverse range of depictions.
But fiction has a habit of romanticizing certain aspects, such as the frequent inclusion of humanoid robots.
Building a robot in pursuit of AI is a bit like constructing the chassis of a car before
the internal combustion engine was even invented.
But I guess anthropomorphism is just an easier solution than expecting audiences to get
emotionally attached to a few lines of code.
Unless you're Joaquin Phoenix of course.
But that dude can do anything.
In recent years a few outspoken intellectuals, such as Elon Musk, Sam Harris, Stephen Hawking, Nick Bostrom
to name a few, have voiced genuine concerns for the rise of artificial intelligence
as brilliant minds across the globe are competing to retrace the hundreds of millions of years
of evolution that resulted in the human brain.
And while many experts have no doubts about machines' ability to achieve human-level intelligence
at some point in the future, just as many believe it to be an impossibility.
And perhaps it is.
Perhaps we are incomprehensibly complex in such a way that we cannot do what nature has done to us.
Then again, if random mutations can lead to intelligence, how hard can it be?
In fact, even though evolution had a hell of a head start
machines have already surpassed us in some limited domains.
Take the two popular board games Chess and Go.
The best human Chess and Go players in the world have been defeated by artificial intelligence.
Just think about that.
No human player will ever again become the best at either of these two games.
No matter how hard you try, even if you dedicate your entire life to become the best player
you could possibly be, it will never be enough.
Luckily for us, there's a lot more to the human condition than Chess and Go but is it
that dubious to think that more general talents and abilities could be rendered equally obsolete?
Assuming it's possible, what happens when machines become more competent at performing
any and all physical and mental labor?
If AI becomes more competent in every regard then what purpose or function would be left for us to serve?
The music playing in the background of this video was entirely composed by a machine.
Here's another example.
How about some acoustic guitar?
Or some drum and bass.
It's quite impressive but, then, take a listen to this.
This short and simple composition may sound rather unimpressive in comparison but then consider this.
...is not the sound of a piano.
The sound itself, the waveform itself, was created by an AI in an effort to mimic the sound of a piano.
Machines have also begun tackling the human language by writing poems.
"When I in dreams behold thy fairest shade Whose shade in dreams doth wake the sleeping morn
The daytime shadow of my love betray'd Lends hideous night to dreaming's faded form."
You know who would love this nonsensical mess of words that is inexplicably a real language?
That's correct, the Danes!
So poems may not be the best example but news articles is a better one.
Machine-written content has become so common place and so good that you've likely read
a few automated articles without even realizing it.
If you pause the video, see if you can discern which of these two paragraphs were written by an AI.
The answer will be in the description.
Now, this is one of my favorites.
Many of you have likely played around with text-to-speech engines.
You input some text and then a voice will read it aloud.
TTS VOICE: Many of you have likely played around with text-to-speech engines.
TTS VOICE: You input some text and then a voice will read it aloud.
TTS VOICE: Please send help.
Up until now, this has been accomplished by having a voice actor read a long series of sentences and words.
Then you splice the syllables into tiny samples before attempting to stitch and morph them back together.
Artificial intelligence can not only do this in seconds but can also produce more natural-sounding speech.
Let's take a listen.
It can also produce human-like gibberish.
[WAVENET GIBBERISH EXAMPLE]
So how does it work?
Well, as you can imagine, it's complicated.
And that's a severe understatement.
But let's use an analogy.
Think of a human infant.
An infant doesn't really know or comprehend anything.
It's just an oblivious sack of meat that occasionally moves around and screams a lot.
Nevertheless, the human brain must be doing something, right?
I mean, it's not dead, so what's happening?
Well, for the most part, the brain will be searching for patterns.
Systematically attempting to identify recurring events in an effort to make sense of the world.
Colorful toys equals fun.
Vegetables, not so fun.
If mommy and daddy can speak, maybe I can speak?
If they can walk around on two feet, maybe I can too?
A few years of this repetitive learning process and we can go from this to this.
And even this.
And unfortunately this.
You don't want that.
This is essentially what machine learning is all about.
You create a simple mathematical model of the human brain and then you feed it a bunch of information.
This artificial neural network will then attempt to make sense of this information by learning
from past mistakes and imitation.
The result is this natural evolution that no human could ever manually program.
And there's really no reason why this system could not be scaled up to human levels and beyond.
A limiting factor of the human brain is that it has to fit inside a cranium.
An artificial brain could be the size of a building or larger still.
An AI could also operate at the speed of light as well as improve upon itself.
In essence, it could teach itself how to learn new things and in doing so would eliminate
the need for a human at the controls.
I don't know about you but the fact that this is even theoretically possible makes me a
bit worried for the future.
There is a genuine cause for concern as lost in this idealistic quest for improvement
is the function of humanity itself.
If we render the species incompetent, what's to prevent artificial intelligence from recognizing
that incompetence and taking control?
Because how do we control something that is more intelligent than us?
It would be like a colony of ants attempting to control even a single person.
It just isn't believable.
And elaborating upon that analogy, think about the way we treat ants.
Must of us don't have some deep hatred for ants but if they are ever in the way of something
that we want to accomplish, we could not care less about their existence.
And even if we did care, how would we explain to them that they need to move because we
want to construct a building or something?
Attempt to explain to the millions of animals living in the jungle the geopolitical and
socioeconomical reasons for why we continue to destroy their home.
They simply lack the intelligence to understand.
And that is what so many are concerned about.
That we would lack the intelligence to keep up.
So this concern for AI has nothing to do with malevolence.
It's simply a fear of losing control.
Then there's the issue of sentience and cognition which are entirely different beasts all together
but if we assume that super intelligence is possible, how long until that happens?
It may be sooner than you think as most estimations seem to land in-between three to six decades.
So within this century.
But I mean it's just estimations. So who knows? No one can predict the future. We don't know, to be honest.
I seriously hope they're wrong though cause it'd be so much more convenient if we could just leave this
impending doom of humanity to our children and just reap the benefits in the meantime.
Happy holidays everyone!