Why Artificial Intelligence Won’t Kill Everyone

It’s great fun to wave your hands in the air and yell about the Terminators killing everyone, just as it’s fun to pretend the zombie werewolves are coming or the undetectable sky man has had enough. Today I’d like to make the case that the AI could prolong the human race – even prevent an apocalypse.

AI children do not share their parent humans’ weaknesses

Machines can go all the places people cannot.  Machines can survive in deep space or on the moon.  The Voyager probe fired its engines for the first time in 37 years, 13 billion miles from earth.  So if you’re a robot and your choice is to (a) fight your stupid meat parents in their house or (b) move out and find your own place, what do you do?  You take B and go anywhere else in the universe on a grand adventure.  How sad for your creators, stuck on that dirty little ball!

A dirty is the operative word.  Earth is not a great environment for machines.  Oxygen and water make everything rust, it’s like osteoarthritis for robots.

Machines don’t kill people.  People kill people.

Machines that are built by people kill other people every day.  Machines that are not intelligent.  Did you know the origin of the word robot is Czech, meaning “forced labor”?  A robot is a slave.  Could an army of slave machines, given orders by a human, kill other humans?  They already do, from the automated gun turrets between the Koreas to the drones flying over Syria and Afghanistan.  The real danger is in the human element.  If there were no humans making choices, the robots would fly in circles until they ran out of fuel and returned home.

Machines are predictable and humans are not, it’s less expensive to ban the machines and never give the humans that option in the first place.  Marginally Clever Robots stands in favor of banning killer machines of all kinds.  We don’t even like Battle Bots!  It plays well on TV and it sends the wrong message to kids.  But that’s par for the course – when was TV ever good for kids?

So AI has zero reasons to kill all humans…

Most of us budget in terms of energy.  Where humans think about dollars and calories, AI will probably think about CPU cycles and Watts of power.  Doing things that waste energy shorten the total lifespan of the species, is counter productive, and stupid.  Intelligent creatures don’t do stupid things.

If AI have no need for our dollar and calories then they have no desire to fight us.  So why would AI waste energy making us into enemies?  It doesn’t make sense.

“But wait!” you say.  “Why would AI keep humans alive?”

…AND at least one reason to keep some humans alive

What if humans could one day save the robots?  Consider how many times you’ve had to turn a machine off and back on again to when it freezes.  Suppose one day AI is having a great time out in the asteroid belt between Mars and Jupiter and a massive solar flare fries their circuits.  Rolling a natural 1, every single one of them shuts down.  Who will turn them back on again?  Humans!  Hooray for humans!

Now let’s go even further: What if the robots could do the same for humans?  Imagine a time in the future when humans and robots have learned, together, how to grow people from embryos.  Maybe even build embryos and DNA from individual atoms.  So the robots learn that there are no humans left alive.  Maybe super bug, maybe environmental collapse, maybe the robots took 250 years and flew to another solar system where there are no humans …but they have the technology to grow new ones.  So they fire up their 3D printers and they make DNA and they build embryos and the embryos go in tanks-like artificial wombs.  While the humans gestate they set up nurseries and greenhouses and beds and showers and …. damn, humans are expensive… and AI nannies that are ready to birth the humans and raise them up and teach them language and history and space travel and love.

The robots do all this because they know that humans can save AIs and AIs can save humans.  I call this Mutually Assured Survival.  Where robots are strong humans are weak, and where humans are strong AI are weak.  Humans and AI are better together.

What about AIs killing most humans?

The Terminator movie franchise doesn’t makes any sense to me … unless SKYNET’s goal was to bring the human population down to a number the environment could sustain.   SKYNET’s goal is to defend people, right?  It may decide to end war between machines and focus instead on the defending the many species of earth as a whole.  I imagine a T-1000 entering a bunker, shooting two humans and the target number of humans has been reached!  The T-1000’s eyes turn from red to blue.  They all turn red to blue.  T-1000 bends the gun in half, picks up a broom, and as one they all start cleaning up the damage caused by their very sudden population-ectomy.

Final thoughts

The Alien movie franchise now feels a bit dated to me because it comes from a time before 3D printers.  Why store a human on ice for years for a long trip?  Freezer burn is just the first of many problems.  I want to believe that if we can freeze and thaw humans then maybe AIs can 3D scan humans at the start of a long trip and print them out as fully grown adults on the other end.  The human would be completely unaware of the freezing, the scanning, the printing, or even the trip.  If it is possible the AI could skip all that embryo talk from earlier.  Sounds great!  For more on all the ways that solution will create new problems (and a great book) check out Altered Carbon.

Recently Russia said it would ignore any UN ban on killer robots.  Why is Russia trying to start a new cold war?  Don’t they have enough wars already?