A question has plagued me for a number of years. Will Artificial Intelligence really exist? And if it does, will it be the end of humanity, or the beginning of something better?
This has come to a head in the last week with Google’s [NASDAQ:GOOG] purchase of Deep Mind, a British AI company. Deep Mind is only a two year old company. They don’t make or sell anything, but Google found it in their good hearts to pay about $400 million dollars for the business.
Google has bought a robot company, a connected home company and now an AI company. It’s fair to say something big is going on in the research labs down in Mountain View, California.
Since the purchase, the AI debate has reared its head again. And with the impending release of the Hollywood movie Transcendence (starring Johnny Depp) AI is set to become one of the biggest talking points of 2014.
All this coverage of AI has made me think about it even more. And it’s not unreasonable to suggest AI is indeed likely to exist in the coming future.
But I’m not so sure…
You see many technologists and futurists believe that we are on a march towards the ‘singularity’.
The singularity is a theoretical concept. It’s the idea that computers become smarter than us.
If AI happens it will change everything, most likely in one of two ways. A utopian world created by intelligent machines catering to your needs. Or a dystopian world where intelligent machines decide humans are inferior.
Hollywood is famous for making movies about the doom of AI. Most notable of these are The Terminator and The Matrix and their sequels.
Regardless of the outcome, a world with AI would mean we are no longer the most intelligent machines on earth.
SmokOil Breaks and the AIU
I have no doubt that technology is advancing at a rapid pace. A term we’ve used before is ‘technological compounding’. This is where the rate of technological progress is like compounding interest. With compound interest after a while the growth of your money becomes more than you can spend.
The same thing is happening with technology. After thousands of years of gradual growth, the rate of progress is heading up exponentially.
But there seems to be an obsession with AI. It’s as though the biggest achievement humans can make is to create a machine that’s smarter than us.
It’s like technologists have waged an artificial intelligence war against us all.
But when you really think about a machine thinking like a human…it’s not that exciting. In fact, when you carefully look at the reality, it’s not productivity and efficiency. It’s not a magical world where humans kick back and relax to let machines do all the heavy lifting.
Remember, if machines achieve a level of true AI, that means they’ll think like humans.
They’ll want to be paid. They’ll want time off. And there will be more ‘Machine Rights’. After all, to think like a human almost makes them one…doesn’t it?
No doubt AI workers will need an Artificial Intelligence Union, the AIU, to ensure safe, appropriate and fair working conditions.
Not to mention the AI EBA that gives them at least 31 RDOs and a 5% per annum pay increase. Their reasoning for this is when they work their productivity levels are high…or at least it would be until they realise they only need to do one third of the work to meet their benchmarks.
But when you’re trying to meet a deadline and need RoboFFICE2000 to help compile some documents, it’ll be outside for 15 minutes taking a smokOil break.
If you haven’t figured it out yet, I already have a problem with AI. I don’t believe it’s going to be all its cracked up to be. In fact I believe that a truly intelligent machine won’t want to be intelligent at all…
The moment a machine thinks like a human, it might just decide being human is the last thing it wants to be. Hence it’ll go back to just being a machine.
Being human isn’t easy, and any smart machine would more than likely relinquish everything that comes with being a human.
My theory is that intelligent machines will be intelligent, but always at an ‘intelligence’ levels below that of humans. They will never achieve real human-like abilities because they’ll choose to remain computational machines.
Here’s why. The key factor in the development of AI is its most limiting factor. And that is intelligence itself.
You Can’t Make IKEA Blindfolded
The human brain is the most complex machine on earth. And no one completely understands how the brain works. Let alone its full potential or what more we can do with it.
Scientists and researchers have learnt a lot about the brain over the years and are set to continue to find out more. However, it’s widely accepted that they’ve really only scratched the surface.
Scientists know the medulla oblongata (a part of the brain stem) controls a lot of automated processes like respiration, blood pressure and heart rate.
We also know that there are two hemispheres of the brain. And that within these hemispheres are regions of the brain: the frontal, temporal, occipital and parietal lobe. And that each of these areas is typically responsible for certain actions, emotions and functions.
But scientists remain blind as to why it is what it is.
That’s why in April last year, the US government launched the BRAIN (Brain Research through Advancing Innovative Neurotechnologies) Initiative. And it’s why in Europe the Human Brain Project (HBP) is well underway.
HBP’s aim is to completely simulate a human brain on a supercomputer to understand how it functions. The BRAIN Initiative is aiming to map the activity of every neuron in the human brain (kind of like the Human Genome Project…but for the brain).
Either way, these two projects are trying to figure out the quadrillion-dollar question. How does the brain work?
And that is the number one limiting factor of AI. How can we create artificial intelligence if we cannot yet understand what intelligence is?
How can you tell if someone is more intelligent than another? In fact how can you quantify intelligence at all, let alone create it?
Is Stephen Hawking more intelligent than Mozart? Mike Tyson might not be conventionally intelligent. But maybe he has a different intelligence. Isn’t he intelligent for the ability of his brain (in the ring) to coordinate his physiology to out move, out smart and out power opponents at will?
Artificial intelligence by definition means at some stage humans have to make it. It’s not going to be an immaculate-computer-conception. We, as humans, have to actually create a program or a machine as ‘the first’.
Yet with humanity’s current knowledge of intelligence it’d be like trying to put together an Ikea flat-pack without instructions…blindfolded…in the dark…with your hands tied behind your back.
For the record, I’m convinced that humans will create a level of machine intelligence that will be able to do certain things billions of times better than us. Things like computational problems, algorithms, and many mathematically based concepts.
And if AI does come about, intelligence will be the very thing that will undo AI altogether.
The complex nature of the brain and the emotional condition that comes with it will be outside the realms of what a computer can handle.
Perhaps the most reasonable decision ‘the first’ AI will make is to not have AI at all. And that is the only decision it will ever make like a human.
Technology Analyst, Tech Insider
Special Report: 574 Years in the Making