Leave a Review
We all use ChatGPT or other AI tools regularly, and many of us have probably felt this:
“AI actually understood what I was saying!”
You explained your situation, the AI understood the context, and sometimes it even gave you a solution that you had not thought of yourself.
So the question is, does AI really understand?
Or is it simply generating responses so well that it feels like it understands us?
This question becomes even more interesting when we talk about AGI.
Suppose an AI solves a scientific problem that humans could not solve before. It comes up with a new idea, writes complex code, or finds a solution that nobody had thought of before.
These can certainly be strong signs of intelligence.
But can we use that to say that AI has consciousness?
In simple words, consciousness means having an experience from your own point of view.
For example, if your hand hurts, you do not simply know that:
“There is an injury in my hand.”
You actually feel the pain.
Now imagine an AI says:
“I am in a lot of pain.”
The question is: does it actually feel that pain?
Or is it simply generating a response that is appropriate for the situation?
We can observe an AI's output, but we cannot directly observe whether there is a genuine subjective experience happening inside it.
And this is where a major unknown remains.
We still do not fully understand how human consciousness works.
Suppose that, in the future, we discover the mechanism behind consciousness.
There was an unknown X, and we finally understand how X works.
But another question would still remain:
Is understanding how X works the same as being able to artificially create X?
One interesting example is light speed and time dilation.
According to Einstein's theory of relativity, when an object moves at a speed very close to the speed of light, time passes differently relative to an observer in another reference frame.
This time dilation is not just a theoretical idea. It has been supported by experimental evidence.
In other words, we understand how this phenomenon works in nature.
But that does not mean we can simply take a spaceship and make it travel arbitrarily close to the speed of light whenever we want.
There is another interesting point.
Because of relativity, traveling at very high speeds can create a kind of "time travel" effect toward the future relative to people who remain behind.
But that does not mean we can simply sit in a machine today and travel back into the past.
So:
Understanding how a phenomenon works and being able to reproduce or control it whenever we want are not necessarily the same thing.
Perhaps consciousness has a similar problem.
In the future, we may understand how consciousness works.
But that does not automatically mean we will be able to create a conscious machine.
Suppose future AI can talk like a human, solve difficult problems, create new ideas, and even talk about its own feelings.
There would still be one important question:
Is it actually experiencing something, or is it simply behaving as if it is experiencing something?
That may be one of the biggest unanswered questions about advanced AI.
Maybe one day we will finally understand the mystery of consciousness.
But then the real challenge may be:
Will we only understand the mechanism of consciousness, or will we actually be able to create a conscious machine?
This is mainly my own observation and thought that I wanted to share. Of course, I may be wrong.
If someone logically analyzes this idea and gives me a follow-up question, a different perspective, or constructive feedback, I would genuinely be happy.
For me, the most interesting part is understanding an idea better through discussion.
And let me know in the comments if you would like to hear more interesting perspectives from me.
If there is enough interest, I will share some more interesting perspectives in the next post. 🙂
And thanks for your time!
