The entire field of AI disagrees with you.
No, it doesn't. In fact some of the foundational members of AI research (and other philosophers of mind) very much agree that what we call "intelligence" is not possible without self-awareness.
AI is already here, and it's all around us: in your washing machine, in your dishwasher, in longshoreman cranes, in your car, in Google, in Facebook etc...
With all due respect, calling that "intelligence" is not very intelligent. Those things are very far from intelligence. It's plain old procedural software, written by humans to do specific things. Yes, sometimes those things include learning, but hell, we have mechanical devices sans electronics that learn. That isn't intelligence either. Nor is your smartphone even remotely "smart"... it only does things that humans programmed it specifically to do.
Both Deep Blue and Watson were essentially "just a look-up program" yet they are considered actual AI, just not the self-aware, generally intelligent kind.
By whom? Certainly not by me, and I am a programmer. Deep Blue was a rule-based game engine. Essentially it generated different "dry run" possibilities until it found a winning strategy. In effect, it was little more than a trial-and-error chess move generator, even though it was very fast at doing so, allowing it to look farther ahead than previous machines.
And Watson is glorified Google Search + Wikipedia. It can search vast stores of information very fast, and make simple pre-programmed inferences from the data it finds. But although it represents the pinnacle of today's "AI" research, it isn't AI. Here's how you can tell it's not even a little bit "intelligent": when it makes mistakes, they tend to be spectacular. Its wrong answers are SO wrong, it is easy to see that it is mainly just echoing existing information, rather than reasoning about it.
So, no: these things aren't "generally accepted" by the experts as AI. The public, which largely doesn't understand how they work, may accept them as such, but that doesn't make it so. Another example: many people still think the Turing Test is a measure of intelligence; however, we now know that programs that have come closest to passing the Turing Test (the real one) are not intelligent at all. They're just programmed to create the illusion of intelligence.