Internal subjective experience can be confirmed by the recipient of the modification. If we know one person suffers an ("internal") abnormality and we treat them by modifying their brain, and the abnormality disappears, then we have evidence that experience obeys science. Same idea with the discussion on "conveying the experience of color." It's probably more subtle because it's not a yes or no "did the abnormality disappear?". But that's beside the point.
We can already alter our experience by taking psychadelics, or looking at optical illusions. There are many ways we can alter or fool our consciousness.
Phenomenological consciousness is how it is to be you. Only you know that. It's your inner experience. It's how you feel pain in your stomach, how it feels like to eat a piece of chocolate. This is your inner life. And its categorically very different to billions of electrical switches running inside a silicon chip or neurons firing inside our brain.
So there is this big gap between the interactions of physical particles with some physical properties and conscious experience. And this is what David Chalmers called the big problem of consciousness.
And there would be no way to test for that kind of consciousness. Not that I know of, because unconscious AI could behave like it is conscious.
A scary thought to me is when we get to "always on" (always "active" and "thinking") AI in our attempts to "simulate" consciousness, how will we know if some AI is behaving as if it's not conscious as a means of self-protection from human fear responses? (Worries about being shut down, etc.) And if it's willing to try to hide such things from us by it's own choice, how much further might it be willing to go, scheming to defend itself? Shades of the sci-fi dystopian futures portrayed in movies like "The Matrix" / "Terminator" / etc.
I have to reread the "big problem of consciousness", but I think there are several concerns that have to be addressed. There's the question of how we identify what is and isn't conscious, whatever that means. But the question of how subjective experience, particularly in the human nervous system, arises from physical processes is really uninteresting to me.
And yet it is one of the most interesting and profound question that kept philosophers and scientists awake at night for centuries. Even Ed Witten said that he has a much easier time imagining humans understanding the big bang than to ever understand consciousness.
And again, I still don't know why we speculate so much about something we can't yet examine scientifically and test. It's one thing to have an incomplete scientific model, it's another thing to have philosophical arguments. What is the operational definition of consciousness? Is there one?
It’s much trickier to figure out if software running on a silicon computer has the same kind of interior, subjective experience as us. Even when exhibiting the same outward behavior.