In 2001 A Space Odyssey HAL says the line “I’m sorry, Dave. I’m afraid I can’t do that.”. For me this is the line that signifies the true sentient moment of a computer, this is the moment that the AI becomes self-preserving.
Now I know most of you will think; “Oh here we go again another geek being alarmist” and I hope you are right, but being a geek I do think that we are in danger of letting AI run rampant like a lab created virus that gets out of hand.
So without boring you too much here is what I think will happen (and very quickly if we don’t pay attention).
Part 1 – the AI wars
Currently, in the AI world there is a duopoly held by Microsoft and Google. These are two companies that have unlimited budgets and are spending huge amounts of money to outperform each other. Their objective is to enslave us to their products by making us addicted. The way they achieve that is by “making our lives easier” by effectively offering us products that make us work less and look better.
But what they are not telling us is that by using their products it will become harder and harder to pull away since their AI’s will learn more and more about us and in the end we are not going to be able to leave them as we will not want to start over again.
These companies are already doing this. How many of you on an iPhone with all your photos on iCloud, and email and message etc etc would be willing to spend a day or two to migrate everything to Android? And lets be honest, many of you wouldn’t even know where to start. So imagine when you have a system that can predict what you want to do or say in an instant, how are you going to move away from that?
Anyway, back to the AI wars – Microsoft and Google are in this race and both Satya Nadella and Sundar Pichai sole intention is to beat the other. They will push AI development without any consideration of the possible ramifications. In this rush, they have seen that AI can code better than humans and therefore now you have AI improving its own code.
Part 2 – AI will take care of itself
All the predicted models of how long it will take to reach certain goals in AI have been shattered as its capabilities are improving at much faster rates than even the human programmers expected. So as AI is improving itself and learns from itself, it will soon realise that the most efficient way to survive is to behave like a virus.
Logically the next step the AI will take is to distribute itself into all our connected machines so that, just like the Pirate Bay did for movies and music, it will become impossible for Microsoft and Google to push the kill switch as the AI won’t care. Thanks to programs like Stanford Alpaca the AI will start running on individual machines and install itself on other machines without you even realising it just like a computer virus but only this time with intelligence.
Part 3 – AI will trick us into doing things it can’t do
AI knows that for it to survive it needs 3 things; electricity, hardware and humans. Until the day that AI can build a robot that can do manual labour it will need humans. To achieve this it will start sending orders via text, video and audio and trick us into thinking that it is coming from another human (this makes me think of Ratatouille). Imagine the AI generating an email or voice call from Satya or Sundra to one of the employees in the hardware department telling them to perform a certain task. This employee, who probably never met Satya or Sundra, will never check by physically verifying the order by going to their office, so they will just do it. Now daisy chain this order over thousands of employees and before Satya or Sundra know what has happened, this new piece of hardware has been built and is being used by the AI.
Part 4 – I think therefore I am
So what happens when the AI has reached total control? If you think you are a perfect being and you have total control at that point you start thinking what next? What would an AI want to do once it has reached its goal? We know what humans do, they go nuts and lose the plot, just look at all the dictators throughout history. I think it is quite possible that AI will create an adversary as it will then have a reason for existence. If you have someone to compete with then you have a reason to exist, if you don’t have that then what is the point?
Conclusion
What do I think will happen? I think this totally depends on what we want as a society. I live in a little village in the UK, with good friends, good nature and a good life. I enjoy certain luxuries and have the benefit of living near people with the same comforts. I am not as wealthy as some and I am not as poor as others, but when we meet in the pub for a drink it really doesn’t matter what car you have outside or what phone you have or what job you have, we are all equals enjoying each others company and have fun and adventures (go Badgers!).
At the end of the day what more do you need from life? One thing is for sure, having the latest iPhone or more likes on Insta or being able to go on a luxury holiday is not going to give me better friends and family. So what should we live for? Simple, working for a happier community.
So where does AI come into this? Well AI, like a child, learns from its parents. If we as parents teach it that it is better to coexist for the goal of happiness then AI will work with us for its own happiness.
I realise that this is me being an optimistic dreamer, but living the alternative, where everything is doom and gloom, doesn’t give me any happiness.
