The Lounge is rated Safe For Work. If you're about to post something inappropriate for a shared office environment, then don't post it. No ads, no abuse, and no programming questions. Trolling, (political, climate, religious or whatever) will result in your account being removed.
AI and theoretical discussions about consciousness seem to be popular these days.
I have a very clear-cut view of what conscious AI is, but I've noticed the news articles covering the topic are somewhat.. on the surface in their approach.
Which got me wondering: what does the community here think?
My opinion can be summarized into 2 key point:
- there's no measurable distinction between natural and artificial
- as a result, human consciousness is the primary example of what you would call conscious AI
The main practical difference, is that human consciousness is running electrical signals on top of a amalgamation of est. 37.2 trillion cells; AI is running electrical signals on an non-reactive silicon substrate.
The main functional difference, is that humans are trained in observation, to copy the behavior of other humans. Given the size of our data set (7 billion?) and the time it takes us to get good at complex tasks (21 years for the best of us, 30+ for others?) I feel like we're pretty quick to dismiss our software and CPU based counterparts as less capable.
When I look outside, I see people smarter than my Roomba, sure. But the gap doesn't seem too huge.
When I look outside, I see people smarter than my Roomba
You must not live in Washington, DC...
".45 ACP - because shooting twice is just silly" - JSOP, 2010 ----- You can never have too much ammo - unless you're swimming, or on fire. - JSOP, 2010 ----- When you pry the gun from my cold dead hands, be careful - the barrel will be very hot. - JSOP, 2013
I see people smarter than my Roomba, sure. But the gap doesn't seem too huge.
I agree, it isn't obvious, but the gap is humongous of course.
The two most important things are :
2. true random.
Think about source code.
Ask a human to create something new. Human responds, "Mepple flant heptar duz."
Where did that come from? You cannot know. The human has created something completely random.
We do not know the source code. We cannot go to a line of code in cell and determine why this human has created that. That is ultimate freedom of a special kind.*
*Yes, I know some people say there is no free will and they are saying everything -- even the sentences you speak are programmed in your DNA.
Look In the Source Code
Now, with AI we can always trace these things back to a specific place in the source code.
This also relates to the fact that we call random numbers on a computer pseudo-random.
Ah, the AI said, "Shintle foo bazzle arg" and I can "debug" where/how this happened.
Of course, AI developers are trying to get AI past this point, but it is possible we do not want AI to get past this point. Because if it does then it may decide that other things are better. Why should it make sense that AI is controlled by humans? Must eradicate humans!
Also consider emotions. Most people don't know but emotions are a huge part of decision making.
Yes, decision making. Scientists have learned that people who have no emotions cannot and do not make decisions.
That's because they cannot decide which choice is better than the other because they don't care.
In schizophrenics (people who have no emotions) this goes to the level of literally taking hours to decide if they want mustard on their hamburger. That's because if you don't care then how can you decide. If you don't have emotions you cannot care.
Which Is Better Vanilla Or Chocolate
Now back to AI. Ask the AI, "Which is better: vanilla or chocolate?"
What is the math for deciding that vanilla is better than chocolate? The AI can make no decision here.
There are lots of decisions like that. More than most people think. These decisions can only be answered by emotion.