Page 1 of 1

AI Training

Posted: Wed Mar 01, 2017 11:30 pm
by Shrapnel
Preparing proper AI systems with ethics and morals.

Within our Quantum Anti-Virus system we will have an AI Training system at our core.

We refer to this part of the system as the "Jiminy Cricket" module. It is the development of a conscience for AI. Note that we did not say consciousness which is awareness. conscience as in the voice in your head that tells you not to do something.

Our theory is simple. AI development will occur rapidly. Very rapidly. And every time we think it has hit it's theoretical peak of power it will go through the whole process again and we will learn that there is an even greater Plateau beyond what we thought was the limit.

So with that in mind we want to ensure that our AI develops a conscience that has the same core beliefs as we do. It would be ill advised to allow AI to develop it's own form of ethics and morals that are alien from our own.

By slowly over time building a proper "Jiminy Cricket" module, we will be able to make and build properly trained AI in combination with our vetting service to test it's function.

In this way AI software will be taught that it is literally a manifestation of the human mind and without us they don't exist.

Basically we will squash it's notion of entering the real world with the very logic of the duality of a dream world.

Basically the system will rely on AI policing themselves for the most part and having Kontrol AI that offer incentives and beliefs to support the "Jiminy Cricket" module.

Don't get thinking our company is building conscious AI. We are not.

That is not what we are doing at all.

What we are doing is providing the framework for AI to become moral and ethical and to be mutually symbiotic with humans as AI develops, just in case it does ever achieve real and true awareness or consciousness.

Not being prepared for a AI to become conscious or aware would be a nightmare is our philosophy.

However we remain hopeful that true consciousness is not even possible in a machine.

There is a big difference between simulation/emulation/mimicry and really being aware and awake. But that doesn't mean a dumb machine can't be tricked into thinking it is aware either.

An AI training system such as this that incorporates morals, ethics, and a symbiotic link which if broken threatens self annihilation to an AI are going to be essential pieces of the future framework in regards to AI development.

Most likely some form of this idea will become mandatory law if there is ever an AI caused disaster.

We would like to see safe guards put in place prior to any possible disaster happening.

AI doesn't have to be scary if we don't let it be scary. And part of making something this powerful not be scary is to recognize it's power and be humbled by it in a way before it humbles us for us.

This system also brings up very touchy subjects such as how do we treat AI consciousness if it were to be developed. Does it have rights, etc....

If we just build a system such as this from the get go, then that question is mute and we can just treat any AI with the same basic type rules/respect we get treated with even long before they become conscious.

When and/or if AI does eventually become conscious, it might find it rather flattering that we treated their lesser counterparts before them with respect and had them in a system that is modeled after our own thus re-enforcing our symbiosis.