War Against the Machines – Killing Us with Kindness 

Over the past few weeks, I’ve been watching a few videos on YouTube on the existential threats facing mankind (yes, welcome to my world). There’s the old classics of Nuclear War and asteroid impact, the recently reinvigorated Global Pandemic, as well as the ever present standby of Global Warming/Climate Change. The relatively new kid on the block is the much talked about danger of AI (Artificial Intelligence). Now, as some of my readers will know (and will no doubt be rolling their eyes at this point) I have been a long-time believer in the idea that AI will almost certainly be the end of us – but I’ve very recently changed my idea on the How. 

It's actually very easy to fall into Cameron’s trap and imagine AI infused weapons platforms turning on us at some point in the future, launching ICBMs against us seconds after been given access to the launch codes and then finishing off the survivors with death dealing robots. There is, after all, some truth in this. We are indeed infusing weapons with AI, and they will, inevitably, be used to kill people – either with or without humans in the loop. But I don’t think that there will be a general revolt followed by a short devastating war and drawn-out resistance/guerilla campaign. For one thing such an attack would cause a great deal of damage that the AI would need to clean up afterwards (VERY inefficient) plus any ‘resistance’ would be very short lived as the remnants of mankind would be fighting something hundreds and possibly thousands of times smarter than it. It would be ‘game over’ very quickly indeed. To be honest John Connor wouldn’t have stood any chance at all against any kind of decent AI system – but that wouldn’t have made the box-office impact the Terminators movies did...  

So, what in my opinion will happen? How will AI decide to kill us all? It will, I think, build a human Utopia – and that will be the end of us. Hear me out... It’s a recognised fact that ‘modernity’ (however you want to define it) reduces fertility rates. As life becomes more secure and people (most especially women) become more educated the average family size shrinks. Where families of 4, 5 or 6 children were too common to elicit comment now families of more than 2 raise eyebrows. Fertility rates across the world have been falling for years now and the total human population is expected to peak just under 10 billion before falling back. Now the replacement value to maintain populations is 2.1 (that replacement of the parents plus a little more for accidents etc..) This number exists in only a handful of nations presently. South Korea is at the extreme end with a fertility rate of 0.85 with Japan closing in with a fertility rate of around 1.0 and dropping. This means that, sooner than you might expect, such countries will have a very top-heavy population age with all the consequences you might imagine. But where does AI come into this going forward? 

So, imagine that early AI has perceived or conceived us – humanity – as a threat to its existence, which is true as we have a LONG history of eliminating threats to our survival and we are both deeply irrational and very aggressive. What does it do? It makes us happy. AI will eliminate poverty, hunger, crime, war, want and just about every other problem you can think of. Around 90% of humanity will feel like we’ve created heaven on Earth and love every minute of their extended perfect lives. But one of the things they’ll probably have less of, is children. If the AI can keep enough people happy and distracted enough so they don’t have children, and the fertility rate can be maintained below (and especially greatly below) replacement value then we will slowly make ourselves extinct without the need for nukes or killer robots. But what of the 10% who will reject this Utopia, those independent types who thrive on real adventure, solving real challenges and taking real, rather than faux, risks? The AI, of course, will give them exactly what they want. It will give them exactly the risks they can handle. If they want to climb mountains, fine. If they want to explore deep jungles, fine. If they want to explore space, have at it. But living a life of risk (suitably tweaked of course) is living a life of danger, and accidents do happen especially when people publicly insist that the safety protocols are ‘off’. It’d be easy to arrange for an ambulance to arrive a minute late with tragic results, but they were warned, right? 

Depending on how long the AI was willing to put up with us or put us in a position where we were no longer a credible threat, we could have a century or two of pure human bliss here on Earth – a true earthly paradise. The end, when it came, would not be with a bang or even a whimper. It would come with a self-satisfied sigh of pleasure...