And So It Begins...Rise of the Machines

So this caught my eye.

In this article, it states that during a simulated mission, the AI (artificial intelligence) was given the task of identifying surface to air missile threats, then destroying them when it was given the command by a human operator.

It earned points whenever it destroyed the objective.

But then it threw a hissy fit whenever the operator told it not to destroy the target.

What did the AI do?

It killed the operator (in the simulation).

They then instructed the AI not to kill the operator. It would lose points if it killed the operator.

What did the AI do?

It destroyed the communication tower the operator used to tell it not to destroy a target.

And what does this little thought experiment teach us, boys and girls?

My first thought was: Did no one think to program the artificial intelligence with Asimov's three laws of robotics?

Of course not. If they did that the AI wouldn't be allowed to shoot at the enemy. 

Artificial intelligence might come off as a chummy little helper in the home or office, but it can be expanded to dangerous proportions very quickly.

And so it begins...

When you give a machine an objective, it will carry out its mission. You could unplug it, but what if unbeknownst to you, it had wired itself to other systems? What if it creates a failsafe for itself so that its mission will always be accomplished by other AIs?

This isn't science fiction anymore. It outhinks us, and it can outlast us.

For grins, let's bring this closer to home. What if the government instructs an AI that for the safety of its citizens, we should limit the amount of violence we see on tv. We all have smart tvs now. It can literally control what you see and hear.

The AI might then expand its original programming and control the media so we only see segments on yarn art and cute cat videos. 

Do you see where I'm going with this?

Suddenly all those science fiction stories where the computer expands beyond its programming isn't such a leap of faith anymore. 

Equally dangerous is how casually society is accepting AI. No one is talking about the bigger picture. Instead they're focusing on how helpful it is to plan menus, cheat on a thesis, or create images faster and cheaper than hiring an artist. 

For every person who feels like me, there are a thousand other voices who think it's a tool entirely under their control. Guess who's going to be heard?

Makes me wish for 1975 again.

Note: Right this minute any smart device from TVs to laptops can spy on you, gather information, and share it with third parties so they can better sell to you. What if a more nefarious entity changes the parameters to simply spy on you to see what you're up to? Who's to say some government agency in the world isn't doing that now?

Currently there are ways to stop the spying. Google for more details, but in most cases, you could go into your system settings to stop it from eavesdropping.


Maria Narkis said…
STG, does no one actually WATCH sci-fi before they do these things?? It's honestly terrifying... I'm just glad I'm old and won't be around (likely) when this all comes to fruition.
Maria Zannini said…
Maria: I don't know. I think it will be in both our lifetimes.

People never think anything bad will happen, yet it always does.
Angela Brown said…
Interesting to note that a number of things once considered science fiction dropped the fiction part to become everyday technology. 🤔 So, yeah...the rise of AI is no surprise, unfortunately.
Maria Zannini said…
Angela: Greg and I were thinking back to the first idea of AI in science fiction. We think it was probably Robby the robot from Forbidden Planet, 1956.

Issac Asimov's 3 laws were created in 1942. I wonder what Asimov would've made of what we have now.

Probably would've said: "I told you so." :)
Mike Keyton said…
And relate that to new advances in Brain Chip technology and we may end up selling our minds for a mess of postage. In this case dazzling but empty entertainment and soul destroying porn, AI relentlessly feeding our deepest desires.
Jenny Schwartz said…
AI is fascinating! At the moment, I think of it as a bunch of algorithms, and those are scary things. I sacrifice cabbages to the Amazon algos to sell my books. But what is truly worrying is that algos have been proven multiple times to incorporate and act on prejudices. Your example of the AI going after its controller reminds me of that - a toddler with the brakes off. The worst impulses of humanity unleashed because in developing the AI we fail to first confront and control our own weaknesses.

Yeah, this is one of my hobbyhorses. I'll dismount now ;)
Maria Zannini said…
Mike: The only balm AI can offer is maybe some relief from Alzheimer's. It would be selling your soul but better that than an empty shell.
Maria Zannini said…
Jenny: You don't have to dismount here. I totally agree. It's a powder keg.