Life Coach

Peter L Nelson PhD

Consciousness & AI

Can intelligent machines be conscious?
I just read an interesting article in Aeon by Prof Michael Graziano of Princeton University entitled: Build a Brain. He states:
"Artificial intelligence is growing more intelligent every year, but we’ve never given our machines consciousness. People once thought that if you made a computer complicated enough it would just sort of ‘wake up’ on its own. But that hasn’t panned out (so far as anyone knows). Apparently, the vital spark has to be deliberately designed into the machine. And so the race is on to figure out what exactly consciousness is and how to build it."

But what is that vital spark? Graziano thinks the theoretical framework that he’s developed called, Attention Schema Theory, holds the key to what has been missing from our understanding of how a machine could become conscious. In the article he conducts a thought experiment in which he guides us from a machine ‘seeing’ a tennis ball, and recognizing it as such, to adding the components that he believes would make the machine a conscious knower.

First he adds a model of ‘self’ to the machine and then a model for the ‘relationship’ between tennis ball and ‘self’. He claims that this would give his machine a model for attention, the
"…mental possession of something, or as something that empowers you to react…Just as my arms and legs are physical parts of me, there’s also a non-physical part of me. It mentally possesses things and allows me to act with respect to those things. It’s my consciousness.’"

Of course, this ability of the machine to answer questions about its knowledge of the ball and its relationship to that ball, still does not address what philosophers have referred to as qualia, or the felt and experienced aspect of knowing. Prof John Searle of UC Berkeley addressed this problem with his Chinese Room thought experiment in which a machine looks like it understands Chinese, but is merely mindlessly looking up characters in a source book and feeding us the translations with no understanding possessed by the machine itself.

Another aspect not addressed in the proposed conscious machine is what philosophers have referred to as ‘reflexive' consciousness: being conscious of being conscious or feeling what it’s like to feel. There is a deep reflexivity of knowing that is experienced by the human ‘machine’ that creates in a human a sense of 'presence' and 'being'. How will we know whether Graziano’s machine has qualia and qualia of qualia?

It could be argued that we don’t really know if we or our fellow humans have on-going experience and that we aren’t just Chinese rooms for each other and to ourselves. However, the immediacy of this knowing we call consciousness and, for some of us, the capacity to know the knowing of others without physical cuing, suggests that something more is going on here.

What that something is, is described by seers in their direct experience of the field ('energetic' presence) of others and what can be known about another through learning how to pay attention to the different qualities and levels of direct experience. Machines can’t do that, especially if you put them in a Faraday shielded light-proof and sound-proof room.

Finally, what have been called 'altered states' of consciousness also represent a conundrum for those attempting to create artificial intelligence that is conscious. More on that in a future blog.