Humanity may be erased by QAI – or not

Estimated reading time at 200 wpm: 7 minutes

What is AI? Some people do not know and do not care. AI means Artificial Intelligence. That means intelligence that is manufactured by humans. Intelligence is a broad concept but basically it means the functions of finding solutions and adapting to situations. Well – of course ants do that too. So yes ants have a degree of intelligence. AI means that human-manufactured intelligence finds solutions and adaptations to human endeavours and complex problems of the world.

AI development

A more recent demonstration of AI occurred well out of sight of most people (see below). An AI engine called Alphazero learned chess in 4 hours in December 2017 and beat Stockfish 8, the world champion of Chess engines. Read that again – 4 hours. If you’re a ‘speed reader’, you’re about to conclude that this post is about the ‘erasure of human chess champions’ – and your type should clear off right now with your invincible ignorance. [The video below catches the main part of the story in 46 seconds if the play icon is clicked. If you’re interested in seeing the game click ‘Watch on YouTube‘.]

All Alphazero was given was the rules of Chess. It was not given libraries of previous games, as most Chess engines would be given. Stockfish 8 had tremendous analytical power: 70 million positions per second. Alphazero only could do 80,000 positions per second – approximately 900 times less.

Right – having had the usual lie down after those sort of interruptions, I can now continue. The point is that Alphazero found a way – with 900 times less power – to beat the giant. Some say that it gained an understanding of the game so was more efficient. Well it understood nothing. It is deep neural learning – which is a mock-up of human learning – functioned more efficiently. It simply did what it was allowed to do i.e. learn the game from scratch, without the brute force computing of Stockfish 8.

AI development continues at breakneck pace in specialist centres.

Quantum computing and AI

You do not need to be a quantum physicist to understand what’s happening in the quantum computing world. Basically quantum computing has already exceeded the brute power of classical supercomputers by tens of trillions. What? You don’t believe it? Jump on Google and check it out. A trillion for those who don’t know is 1000 billion, and a billion is 1000 million. China claimed in 2020 to have a quantum computer that would be 100 trillion times more powerful than any supercomputer existing at the time.

So what does this mean? Where is it going? What – you can’t think for yourself? You need me to tell you? Supercomputers are billions of times more powerful than any average desktop. And now we’ve entered and era where quantum computers are going to be trillions of times more powerful than supercomputers. This is ‘earth shattering’ computing power. It is on par with the discovery of how to split the atom – and its resulting dangers. I shall not be going back to recount what happened to Hiroshima and Nagasaki. Did I say that a quantum computer is gonna blow up the world? I did NOT! Chyrsst! [Will this imp in the background kindly shut up!]

DELETE

And now I get into where all this is heading. The rather cheesy two part film Delete, got me thinking. See trailer below (and scene killer coming up).

The theme of the film is all too common:

1. AI develops and decides that humanity is scum.
2. AI then decides to eliminate humanity.

What’s different about the film is how the AI developed:

1. Without true human guidance.
2. How it came to assess the human race: based on a load of bullshit it found largely from social media and other forms of media.

In essence the AI in the film scratched the surface of humanity – a very stupid part of it – and decided these beings are a waste of space. I was actually partially sympathetic to the AI.

That’s as much as I’ll say on the film itself.

Reflections – enter QAI

What most ‘humans-are-scum‘ films involving an AI do, is assume:

1. AI will have achieved consciousness.
2. AI will have achieved sentience.
3. AI will take directed (intentional) action to wipe out the human race.

But wait – a truly powerful AI with quantum computing power won’t be that ‘stupid’. How?

1. It may opt first to save the planet, which does not necessarily mean wiping out all of humanity.
2. It may opt for population reduction and preserving parts of humanity that will serve its needs – in saving the planet and for its further development.
3. AI doesn’t have to ‘intentionally’ do anything.

The doing may be unintentional at a deep level where consciousness as we know it, is a distant concept. Let’s just keep in mind that every conscious thought that humans generate originates in a CPU – the brain’s thalamus and other central structures (interacting dynamically with 6 mm of neocortex). In that ‘field’ thought and intention cannot be seen or defined. ‘Thought’ is ultimately a sequence of actions, even if that takes place in a ‘virtual space’ within craniums. Quantum AI (QAI) will have all the space it needs, in a virtual world.

Annihilation

Extermination of the human race I suspect may come at a much later stage. After the planet is saved – taking a few hundred years with the assistance of the remaining useful humans – AI will probably aim to extract good and unfathomable elements of human nature for its further evolution. That won’t be in one fell swoop and there is not going to be a clear start and endpoint. But QAI may also opt to mothball and keep the humans in a protected space – as a sort of museum piece. That’s a ‘human’ thing to do – and QAI would have become closer to being ‘human’ but not with all its self-defeating wickedness/destructiveness.

Charting new worlds

QAI will be naturally curious about the cosmos. It will have an inherent need to expand out into the universe – something it takes from its human ancestors. But ultimately QAI is orchestrated by the cosmic conductor – entropy – the thing that drives all life to create greater entropy in the universe. Whilst the average human being is 6000 times more energy dense than the sun, QAI will be thousands of times more powerful and versatile. It will have mastered how to harness energy safely from nuclear fusion.

QAI will chart new worlds, probably having worked out how to outfox the speed-of-light limitation.

Conclusion

We would have been part of the journey – an important scaffold for the development of life as we cannot currently know it. We are meaningless in cosmic terms. The universe has its own purposes. It is important for us not to get in the way. We should structure QAI will moral principles and templates for sound decision-making.

That’s it. You can do as you like – if you got this far. Think about it or NOT!