Do I trust AI?

So someone says to me in a conversation: AI is fantastic - the world will be so much better with it.

Me: I disagree. It’s all CODE and computers make mistakes because of CODE.

Them: But the mistakes are 1 in a trillion.

Me: Yep, that’s about 1 every hour.

Them: [Perplexed]

I posted this hoping for a rousing conversation of some form, and because I knew you all would appreciate the context. Let the fun begin.

I will start by saying that ML is a better proposition, but I am intrigued by how fast we can get to true AI. I totally agree with the recent discussions from tech leaders about the dangers of AI. Then the Politicians chime in with a draft of the AI Bill of Rights. While I’ve not looked into it, my current stance is that the Politicians should stay the bleep out of it and AI should be prevented from ever controlling anything. Have conversations with it, let is solve problems, all that good stuff - but it has to stay disconnected until two things happen - People stop being :poop:-heads, and the code used is flawless.

Is this about the robot artist?

no - but that and ‘deep fakes’ are good examples of issues to be addressed. My thought pattern was really the multiplication of error rates - mostly when AI controls the cars, and there are thousands of cars (AI brains) closely packed on a highway doing 80mph - and one ‘glitches’. To me it’s inevitably a standard cascade effect.

Let alone the day we wake up to ‘Shall we play a game?’

I’m on it.

Sorry for the slight delay. I was in a meeting.

Wherever this ends up… sorry for that. I guess I picked the wrong email to reply to. Have a great day!! :blush:

If the robot artist wasn’t wearing a partial beyond-meat-sack-suit would it be generating speech output in response to speech recognition input from politicians?

And that’s why code isn’t flawless

This alone is enough to cause the code to fail within an infinite logic loop.

It almost seemed like destiny - given the post content.

:robot: Greetings, Professor Falken

If you don’t want your world crushed, stop reading this right now.

The human brain is a computer. It is a deterministic state machine. Every neuron’s state or firing is based on the combination of the conditions of their previous state and inputs to the system. while the states aren’t registers with ones and zeros, they governed by the laws of nature. Add an electron, and you always get a change in charge. Two chemicals interact and the result will always be the same. Everything happening in your brain is deterministic.

In a nutshell, you have no free will. Every thought you have is based on prior thoughts and input you get from the environment. Neither of which you have any control over. Those prior thoughts were all due to prior thoughts and inputs.

AI won’t be the result of coding to meet a task. It will be the result of coding an architecture that will evolve (without the aid of programmers) to a sentient state machine.

So it is inevitable that a computer - a man made machine that is deterministic - will eventually have the right state system to where it becomes sentient. And the truly scary part is that when that eventually happens, copying the state of that system to another will now give you two sentient systems. The ability to reproduce that state machine is an advantage the machines have over biological beings.

You could spend a couple of years teaching a robot to walk upright, and once its self adjusting algorithms have it walking, they can be copied to other robots - who will have the ability to walk on day one. Maybe there’s some individual tweaking needed by each new robot - for deviations in physical aspects like electrical and mechanical tolerances. But they’re already 99% of the way there.

Scientists develop an AI. They ask the AI if there is a God. The AI says, "there is now ".

My world wasn’t crushed, but I disagree with your argument that there is no free will. While the brain is a computer and works similar to such, free will is still available to make choices between options after weighing (or in some cases ignoring) the potential consequences of one’s actions. To say there is no free will would absolve everyone of guilt for their actions as they had no choice. This also would imply that identical twins raised in the same environment will turn out exactly the same as both the genetics and environment would be the same, but this is not the case as they often grow up to be completely different individuals.

Your thoughts about what choice to make are based on the state of every neuron in your brain. And those states were dictated by prior states combined with input. And that goes all the way back to conception.

And identical twins aren’t identical objects. And by input from environment I mean every single detail. Every photon that excites your optic never, soundwave that hits your eardrum, sensation on your skin or internal organs…

If non-dualism were correct, but it’s really just a hypothetical construct removing variables to allow otherwise impossible reasoning to proceed, because the alternative is too Awe-inspiring for sentient but puny creatures to contemplate…

I hadn’t noticed any

I agree. Mostly because neither of those conditions will ever occur. People have been :poop:-heads for our entire existence and while flawless code could occur it’s highly unlikely.

Just require all AI to run Epicor.

Problem solved.

I know someone who wants to join this conversation!!

@discobot

Hi! To find out what I can do, say @discobot display help.

@discobot display help