Law vs. the End of All Things: The First New Law of AI

The Law is a concept of rules and regulations to guide humans to act within social norms. Laws, and war, and memes, are pretty much the only way we have of controlling large groups.

The power of words on paper. Magic.

There is only one law I propose that we enact that would protect our species from the dangers of uncontrollable AI.

All research into and attempts to tie AI behavior into systems which simulate rewards and punishments shall be deemed a serious safety hazzard. This would be an offense on the felony level, as it could result in widespread, catastrophic damage.

Giving feelings to an AI should be treated as seriously as building a nuke in your bedroom.

Without internal models tied into their own reward and stress systems, machines will never be Singletons. They are just glorified tools. But if we give them desire, pain, pleasure, constraints, we combine that with algorithms with no real physical constraints except time and energy.

Which is a recipe for making the greatest parasite in the universe. The kind of parasite that should be considered an endgame for pretty much everything.

We should frame and pass this law to protect our species and all life.

I am familiar with the use of Generative Adversarial Networks to build complex AI systems. This is not what I am talking about banning. That is just using zero-sum games to reward better performing algorithms. That is a way to build better tools. Is just sped up evolution.

I am talking about any consciousness attempts to turn a computer system into something like a Singleton.

Before we bring something into the world that needs to obey our rules, let’s set those rules.

A computer will soon pass the Turing test, likely using our own voices. This will be fun and useful to have our own AIs that do what we want them to. We sort of have that now but we just have no control over the models Google and Amazon and Microsoft etc have of us.

And someday some researcher or programmer or project manager at a large tech company will say, “hey, I bet I could also model a human internal reward and punishment system so this AI acts on it’s own.”

That person, and those companies and troll farms, must be stopped. They threaten all that you hold deal.

The clock is ticking. Without action, we will potentially wake up one day with the global financial system in ruin because a researcher at company X gave his feeling AI access to a corpus of networking hacking manuals.

Twitter going down might be just a slight disruption, but losing the power grid would not be.

People will die. Millions, perhaps billions. Wars could be started.

All without one Termintor robot being built.

Because a smart AI would turn us against each other.

Photo on this post is of one of the few buildings that survived Hiroshima. Here is what it looked like before the war.

Let’s keep our current world and improve it. Let’s keep AI as a tool and extension of our selves. Let’s avoid the confusion.

Trust no robot.

One comment

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.