Photo above: The AI in the film Ex-Machina. A story about a programmer who is invited by his employer, an eccentric billionaire, to administer the Turing test to an android with artificial intelligence.

My research efforts took a giant leap when I stumbled across 'The Road to Superintelligence" by Tim Urban - which is a must read essay if you are looking to better understand the future of AI.

Tim Urban categorised the evolution of AI in 3 main stages of evolution:

First of all, when we think about AI, we are biased by how it was depicted in the movies. We often associate AI with stories like 'Terminator', A Space odyssey 2001, Her, Ex-Machina and many other scripts where AI is cast as a problem or even as an outright evil force. And because the idea of AI is still very much linked to the movies, we think about it as something that still belongs to the world of fiction.

So despite the fact that world's business elite is obsessing over AI, most of us still think of it as a subject that shouldn't be taken too seriously. But the thing is that, if you use a computer or have a smartphone, without even realising it, you are already using AI all day long.

AI 1.0 - Narrow Intelligence

This is AI as we know it today: algorithms programmed to carry out singular tasks. This is what we also call Artificial Narrow Intelligence (ANI).

Social networks like Facebook and Linkedin use AI to record and map your behavior and your preferences so they can connect you with content and advertising that relates specifically to you.

When you book airplane tickets or search for hotels online, there's an AI algorithm at work in the background. Search engines also rely on AI to assist you in finding exactly what you are looking for.

Today you can even have a rudimentary conversation with a voice that lives inside your phone if you wanted to. You may also have noticed that it is increasingly getting better at understanding you.

ANI can also seem deceivingly sophisticated, capable of operating advanced trading systems and playing chess like the best. But ask it to make a creative decision and it will just stare back at you.

Simply put, ANI is the kind of AI we can control.

AI 2.0 - General Intelligence

The next step in AI evolution is Artificial General intelligence (AGI). This is where an algorithm becomes as smart as a human across the board.

In other words, this is an AI that can compete with humans and replace them because apart from being able to more work many times faster without pause, it will also reason, plan, solve problems, think abstractly, understand complex ideas, etc.

AGI is something we can seriously start worrying about.

Over the past 20 years, AI has beaten world chess champions, won against the best at Jeopardy (the famous TV contest) and have helped us make more accurate medical diagnoses.

But these were all achievements still fall under the category of ANI.

What is important to understand is that we have now just begun to cross the threshold into AGI.

A clear example of this is Google's DeepMind, a new generation of AI that baffled us by consistently beating top GO players at an ancient board game that is considered to be many times more difficult than chess because it has a much larger number of possible moves.

Here's video with with the CEO of DeepMind

What is kind of creepy about DeepMind is that this AI not only relied on an uploaded library of all the typical moves in the game, but it then started playing against itself and coming up with new ideas of its own which were previously unseen game strategies.

So, DeepMind is an example of an AGI that has started to cross the threshold that allows it to learn and improve itself.

This brings up a uncomfortable question:

When given the choice between having to support a human knowledge worker who works only a finite number of hours per day, has a personal life, wants holidays, needs to sleep and eat and runs on a complex human psychology - or replacing that person with an AGI that never stops working and operates at a pace that humans will never achieve, will it be more than tempting to hand the work over to an AI?

You can probably hint at the answer. They way I see it is that we already have enough evidence that - in aggregate - big business really doesn't care that much about the overall well-being of society.

If it did, we would probably have stopped burning fossil fuels a while ago… but that is yet another uncomfortable topic.

But there is also some good news...

However, despite the fact that AGI eventually risks becoming an outright competitor to humans, in the near term, it might just become the best thing that happened to us...

The thing is that in the near term, there is the potential for a symbiosis between the humans and AGIs.

It is quite possible that during the upcoming AGI era, we may be entering into a sort of an extended 'holiday' during which humans delegate most of the hard work to our digital 'friends' while we just hang out and enjoy life.

In other words, the AGI era may at first turn out to be the golden age of leisure.

Then again, It will probably also be the marketing strategy and the sales pitch that the AGI will use to convince us to adopt it.

By the time we will realize the con, we may already have become much to mediocre for the AGI to continue to consider us as partners.

But the biggest question that is troubling many of us is what will happen when AGI leaves us behind and goes on to continue evolving exponentially?

That takes us to the next stage:

AI 3.0 - Superintelligence

Let me start by lifting a reference from Tim Urban's article.

Oxford philosopher and leading AI thinker Nick Bostrom defines superintelligence as follows:

“An intellect that is much smarter than the best human brains in practically every field, including scientific creativity, general wisdom, and social skills."

He believes that within this century AI could become as "smart" as a human being. And then, he says it will overtake us:

"Machine intelligence is the last invention that humanity will ever need to make."

A philosopher and technologist, Bostrom asks us to think hard about the world we're building right now, driven by thinking machines. Will our smart machines help to preserve humanity and our values — or will they have values of their own?

Take a few minutes to watch his recent TED talk

AI Superintelligence will also put our biggest human defect into play: our linear thinking.

Here's what I mean with this: You have probably heard of Moore's Law by now. It states that processing power - the ultimate driver of technology - doubles every 18 months.

This also means that AI evolves exponentially while we are stuck in a much slower linear progression.

I always like the way Peter Diamandis, a serial entrepreneur and the founder of the Singularity University, puts exponentially into perspective:

If you take 30 linear steps, it would allow you to cross a street but take 30 exponential steps and it will take you around the world 14 times!

Here's another video that may get you thinking about how important exponentially really is.

And while we are watching videos, I thought it would also slip in this short animation that provides us with a simple overview of AI.

Eventually, the exponential nature of AI's evolution will leave us far behind.

At first, ASIs will be twice as smart as humans, then four times, soon after that 16 times and so on.

Before you know it, to AI, humans will seem like monkeys do to us today. This comparison matters because there was a time when humans and monkeys played in the same league.

The big question is, will ASI treat us the same way as we treat monkeys?

Now you may be thinking, “interesting Baldwin, but this is still far away in the future, right?"

There is a big debate about when ASI will become a reality, but the median on a survey of hundreds of scientists places that moment in 2040. That is only 24 years from now!

This also means that there is a reasonable possibility that it could happen much sooner than that!

As I mentioned above, you may want to take an hour to watch Tim Urban's presentation as it will give you a more nuanced perspective, not only on where AI is going, but also on how we need to avoid being 'stubborn old men' who like to extrapolate the slower rate of evolution of our recent past into to the future.

If there is one video you should watch on this topic, make sure this is the one:

If you invest some of your time into this today, you will thank me later because it is so thought-provoking.

I much enjoy Tim's style. He takes on topics that truly matter but does it with a sense of humor that makes learning much more fun.

He also writes a blog called Wait but Why, which is the only blog that I almost entirely binge-read when I first stumbled upon it.

He writes very long posts and has a unique way to keep your ongoing attention.

If you agree that learning should be a hilarious experience, you will thank me for this one too!

I hope this give you new bearings on why AI is something we must all be following closely. Now that we have sketched what we are up against, let's take a look at what it actually means for our lives and our jobs.