in other words he is essentially an illusion of AI rather than an actual AI.
The term "AI" doesn't mean it has to be a self-aware consciousness. He is an AI, and a rather good one at that. However, he is not a "strong AI", which is what you are talking about. But that doesn't mean it's impossible to design a machine with strong AI. I think it is possible and if we're lucky we will get to see it within our life time. Your point about having the ability to understand abstract concepts and develop new concepts is right on the money. When we have a machine which doesn't just regurgitate scripted phrases, and instead puts together its own phrases based on internal conceptual models which are dynamically updated by the machine as it learns new information, then we will be on the path to having true strong AI.
The term "Artificial intelligence" contains the word "intelligence" which is perhaps not the best word to use to describe that machine then.
It now appears that Facebook has joined Google in the race of AI, am I the only one who is tired of all the psuedo-intellectuals nerdgasming in interviews with their wild imagination which doesn't have any basis in reality?
The real headlines should read "Facebook has joined Google in the race of propaganda" because that's what it is, a non biological computer never has and never will achieve an intelligence higher of a cockroach, let alone of a human being capable of reason and thought, it reminds me of Apple and their "revolutionary technologies" as well as of the sheer ignorance of the crowd as they applaud a robot that has been scripted to perform multiple sequence of moves by a bunch of chinese liars.
Do you believe in a future predicted by sci-fi? Do you believe that one day you could philosophize with a robot?
Assume for a moment that we can make artificial neurons that would be functionally identical to natural ones, and then we begin to gradually substitute neurons in brain with them. If consciousness would be preserved, then yes, the answer to your question is positive...
It would be hard enough to simulate a single neuron, let alone billions of them on the same scale is impossible (and illogical).