I don't think that Facebook's facial recognition algorithm is conscious, but then I don't think that humans are either. Daniel Dennett's definition of consciousness, that you state, is useless because it simply gives an extremely vague "explanation" of how this undefinable thing arises; it doesn't define it.

A piece is missing here: There is no "cognition process" what so ever in neither a camera nor an algorithm.

No, there is a similar process in deep learning (DL) algorithms to what happens in the brain. I posit that the only difference is that the brain has pattern recognition systems that imagine a self.

A DL classifier is not a camera; it is a system that you can show an image to and it will say, "cat," or "dog," or "your great uncle." It's an algorithmic implementation that mimics the hardware/software functionality of the brain, that mimics the brain's cognition. This is why we now have DL systems that are as intelligent as, or more intelligent in some ways than, small creatures.

My argument is that the only missing component to create "self" is a set of these DL systems that introspect. No big deal.

Thus, data engineers play God or Frankenstein when they believe that their algorithms will eventually make their machine alive.

You play God when you think that you're "alive." There's nobody in there. There's only God.