Will we need Artificial Consciousness before we can create true AGI?


#1

In terms of true creativity and being able to adapt to situations where there is a lack of data, fuzzy rules and/or fuzzy goals… could an AI really do this enough without some level of ‘free will’ and a conscious self awareness to really be good enough to be called a true AGI?


#2

I suppose we’ll find out. lol :slight_smile:


#3

Artificial consciousness in terms of a legal/moral framework, would be a useful guide for uncharted situations.


#4

Hi Mike. I hope you know you share the name of the AI in the Moon is a harsh mistress by Heinlein. And the answer is pretty similar to the book. Most of what people and animals do is just fully on auto-pilot and is a consequence of perception of the environment interacting with basic drives. The AI bit is the creativity and planning part, but even those are just higher level automatic things since creativity is 99% about memory and analogy i.e. solving or designing by adapting existing solutions, and the last 1% is a combination of visualisation and noise and filtering / biasing by memory and personality. The last bit of AI is consciousness and idea of self. Every robotic AI will model itself and so in that way be conscious. However, the critical distinction is the hardware on which this is running. An AI is running on reconfigurable electronics that changes purpose moment by moment. A biological AI (and I do mean AI here) is running on dedicated circuits for every function and so the state of the system is preserved from moment to moment and interacts in a highly analog feedback loop, think of a microphone and a speaker feedback loop creating this chaotic and dynamic sound. This moment by moment analog correlation is what I believe creates the FEELING of consciousness, but it is not necessary for intelligent behaviour to have this feeling i.e. it is a byproduct that in itself is not useful. So, until we learn to grow 3D circuits that are dedicated to each AI function then an AI device is in fact being SIMULATED and its results are useful, and with the 3D grown dedicated device then truly we have an intelligent thing with consciousness. That I believe is the only difference between what we call A(G)I and actual intelligence, simulation VS native running.


#5

Very interesting idea on consciousness, thanks for the answer! :slight_smile:


#6

You are welcome. I have worked in practical commercial real-time AI and computer vision for 32 years since I was 20 after uni so I worked out quite a few hypotheses over this time. One of these is the role and origin of conciousness which I am guessing is this dynamic effect of dedicated temporally correlated “hardware”. I could be wrong of course :slight_smile: !


#7

I do believe that is a necessary component. The question then becomes, where is the line drawn between artificial consciousness and just plain consciousness?


#8

Very true. Will be truly bizarre to see the huge ethical & philosophical questions we’re going to have to face once we’re in that position.